Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
/platform engineering: game-changing discipline
58 Plays4 months ago
In the latest The Forward Slash podcast episode, we’re diving deep into this game-changing discipline with Joel Vasallo, Senior Director of Platform Engineering at TAG - The Aspen Group. From redefining workflows to tackling unique challenges in healthcare and beyond, this conversation is a must-listen for anyone passionate about building better systems and better developer experiences.
Recommended
Transcript
Importance of Education and Training
00:00:05
Speaker
It starts with a lot of education and training um and and kind of understanding as to what the platform can and can't do. Out of the box, you know, the way we see is we build a platform with empathy for our developers. If a developer can cause a systematic break, um it's a fault to the platform, not the developer.
Introduction to Platform Engineering with Joel Vasallo
00:00:28
Speaker
Welcome to the forward slash. I'm your host Aaron Chesney with my beautiful co-host James Carmen. Today we're going to be talking about platform engineering, what it is, what it's for, and our special guest who's well versed in the subject, Joel Vasallo. Joel is a senior director of platform engineering at Tag. And I don't mean the, the playground and recess game that everybody plays. It's the Aspen group.
00:00:54
Speaker
There he covers cloud DevOps delivery and SRE strategy. Also a Google developer but group organizer for Chicago running monthly meetups and events to help grow the Chicago tech seat. Welcome to the podcast, Joel. Thanks for having me, Aaron. Thanks for having me, James. Happy to be here.
Two Truths and a Lie with Joel Vasallo
00:01:13
Speaker
So Joel, to let us know a little bit more about you, why don't you give us ah two truths and a lie and we'll see if we can't guess which one the truth is.
00:01:24
Speaker
All right. Let's see. Here we go. Um, like I always think, I always think of these. so Let's see if we get them. Um, I've broken multiple Lindsay concerts. Uh, I've run cube CTL against the production system from my laptop before, uh, or I work out every day. Well, I know which one's a lie about me. yeah Same. Same.
00:01:51
Speaker
Actually, I think all three of those are a lie about me. Oh, I've definitely done the second one. Sometimes you got it. Sometimes you got it. Yeah. Uh, I don't know, James, which one do you think the lie is? Uh, well, um, I don't know. That's, that's a, that's a tough one. Breaking limbs at a concert sounds a little,
00:02:17
Speaker
crazy like that sounds far-fetched but I don't know i didn't I don't know Joel very well so maybe maybe that could be I don't I don't know maybe maybe we leave it as a mystery to when I'm not Kubernetes ing it I'm uh I'm throwing it down in the mosh pit is basically what I'm throwing down that's right Or, I mean, if you said I got my limbs broken because I ran cube CTL on a production cluster. wow that could not yeah believe that that I mean, I've done some crazy things in a production system before. So I'm, I'm definitely thinking that the cube CTL things is a truth.
00:02:52
Speaker
no ah but we're all for people we know yeah that That's not that far of a reach. Um, So I'm between the broken limbs, at broken limbs at a concert. I mean, mosh pitting, that's pretty violent. I could see that working out every day. That's, that's a lot of dedication. i mean I mean, my son works out a lot and it's not every day, but you know, once he goes to that third day, he's like, I got to get to the gym. I'm like, okay, whatever. I don't know how you're my kid, but, um, so I, I think I'm going to go with the, uh, the workout every day.
00:03:30
Speaker
as, as the lie. Yeah. Yeah. I mean, I think that's, you're actually right. Uh, I don't work every day. There's a thing called Rust days. So I wouldn't be able to keep CTO off my laptop every day.
Understanding Site Reliability Engineering (SRE)
00:03:42
Speaker
ah So Joel, we, it was mentioned in your, in your intro that, uh, your SRE as part of your title, so, or or as part of what you do, what is SRE wouldn't explain that a bit.
00:03:56
Speaker
Yeah, um it's it's it's a pretty ah it's starting kind of Google's book on SRE. um and I mean, i I always recommend people to new to this space to kind of check that out. But basically, ah SRE, fundamentally speaking to me is ah you're monitoring, you're learning and you your ultimate incident management strategy. um So think about it, things break all the time, right? But all that really matters is how you respond to it. Do you have a work around? Do you know how to visualize the the impact and also the recovery? You know, oftentimes when things break, we say, hey, we got to fix, we roll it out.
00:04:31
Speaker
so Is it really fixed? Is it still down? um The SRE team that we kind of oversee and and my teams oversee um really focus on um incident management as well as ah basically are kind of the arbiters of our infrastructure. um They kind of have a pulse on the and everything that kind of goes down. They work hand in hand with our application owners, you know whether it be something as simple as Microsoft Teams or something as complex as a database and an application stack. um They basically work to find out what how and how we can monitor systems as well. um You know, sometimes it's hard to think about that, but it's it's ah it's it's definitely a lot of work, but it's very, very cool to see once you have it implemented. So what does SRE stand for? Yeah, ah Site Reliability Engineering. So okay i was kind of keeping in that.
00:05:24
Speaker
It was thinking system recovery something or system recovery engineering. So I was off a little bit. You're not off and the the light so because half the time you're recovering something. each year But ah no, yeah, I mean, it's there's there's a whole established principle around like metrics and monitoring, but even like response times, ah you know, we hear things like what's your SLA and what's your SLO and how long are you up and
Industry Standards: SLAs and SLOs
00:05:48
Speaker
down for? Right. I mean, the many of us in the industry we hold our vendors and we even hold ourselves to these numbers right we're five nines well does everyone understand what that actually means in terms of how long you can actually be down um so i think that's something in the five nines is five nines past the decimal point right so it's 99.99 no it's just all it's all livess you know so it's 99.999
00:06:14
Speaker
so Yeah, thats so um if you say you're five nines, you should expect less than six minutes of downtime over a course of a year. So easy to say, ah but as all of us know, hard to do. that's That's a tough one. thats it' tough Yeah, I don't I don't think people realize, you know,
00:06:29
Speaker
you know, the timeframe that that means to get when you add just another nine. Cause I know it was like, you know, we, we talked about like, you know, okay, we're, we want to be three nines, four nines. And you know, every time you add a nine, that time gets so small. It's ridiculous. He used to be a, well, we can be back up in a day.
00:06:54
Speaker
When are, when are the next Olympics? We'll be back by then. or I remember we used to do like full day shutdowns for, you know, service changeovers and stuff like that. It was, it was like, okay, well we're going to have a maintenance window of, you know, from six to six so that we could get our stuff done.
00:07:14
Speaker
Well, I mean, it's a sign of a, it's a sign of maturity in the, in the technology that we use. We are dependent on it. I mean, even our heck, our cars need it. Our microwave. And you know, I saw someone's fridge needing a software update the other day over the holidays. I was like, your fridge needs to update. It's like, one of those smart fridges that.
00:07:34
Speaker
It will update. i'll hear it I'll hear it update itself every now and then I'll hear the the little chimey thing as it boots back up. This is awesome.
00:07:44
Speaker
a ah It basically comes down to anything that connects to your wireless network has the ability to self update now. So it's it's kind of strange.
00:07:58
Speaker
It's weird. i and We've talked about it. We've talked about like the personal assistant things with like, I'm going to set off everybody's stuff with like Siri and Alexa and and all that stuff and in Google. And it's just, it's, it's weird sometimes how They'll just go off on their own. Like we've been having a conversation in the room and Alexa just butts in and it's like, we're not talking to you Alexa.
00:08:29
Speaker
but maybe you want to yeah Yeah. So yeah, it's, it's very interesting. So with now. With the monitoring, obviously that's part of do the DevOps flow, right? That last part is the monitoring and feedback that feeds, you know, feeds back into the development cycles. You know, how, how does that map with like platform, platform delivery and your cloud teams and, and, and all that, like how, how do you integrate all that stuff in?
Developer Ease and Platform Support
00:09:03
Speaker
Yeah, I mean, it it kind of to the flow that you talked about with DevOps, right, as a whole, um it's part of that experience. um And it kind of goes to what really a platform is. um You know, a lot of us in the industry nowadays, like say, Oh, we're platform engineering, so that must mean you must run a backstage or you know ah a port or a cortex or something. That's just a means to an end. ah right like That's a tool and a platform for visibility. But really, a platform is what you build for your internal teams um or your developers to kind of give them that warm and fuzzies. right ah like as ah Think about it like as you start your first development job. right
00:09:45
Speaker
You got to set up your laptop, you got to set up your Git repos, you got to set up all this stuff. Really, if your platform is built with developers in mind, day one, they should have everything that they need to kind of get started, um right? ah How do I make a repo? The platform can help you with that. Oh, cool. How do I update an existing repo? The platform should be able to help you with that.
00:10:04
Speaker
I want to increase the memory of my CPU and memory and CPU on my deployment. The platform should be able to help you with that somehow. um So kind of bringing all that into the flow um and kind of tying SMEs and the subject matter experts into it. If you build like a stable platform, everyone could kind of influence different components of it. It's kind of how we try to do it. Yeah, and I actually do as as part of our onboarding at Clarity.
00:10:32
Speaker
um I do a talk on DevOps and a lot of the misconceptions is that it's a person that goes out and and gives you all the tools and and does deployments and and that kind of stuff. And it's like CICD and all this stuff. And it's like, no, DevOps is is it's a mindset, right? it's It's dev and operations put together. You should be able to take your code from inception to running on a system in an environment that that is integrated with the ah whole
00:11:05
Speaker
the whole story, right? It's like your chapter fits in the whole story and you can see it running and then you can monitor it and get your feedback and all done as, as a personal mindset. It's like your job doesn't end when you check your code in and it builds, right? Amen. Right. It has to run.
00:11:27
Speaker
um And it has to run in environment. So something else takes care of that Aaron some we have other people for that. Yeah that was very or I mean that's the way it used to be for us I mean it used to be that you had developers and you had operations and you had to make really good friends with the people that had root password Because that was the that was the only way that you were gonna get your stuff fixed, sure but it ran on my laptop and Right. Yeah. Yeah. I mean, that's where it runs on my machine. You know, that was the, that was the the phrase, right? Well, random my machine while the systems are pristine. Well, I don't know what to tell you.
00:12:03
Speaker
And then, ah hey, my, my container ran locally. then oh yeah right That's the newer version, right? exactly yeah random my container I mean, I think like in all seriousness, like, I mean, that in itself drives kind of the, uh, the conversation, right?
00:12:23
Speaker
What is the contract to get onto a platform and really for us is kind of joking about is you have to have a working container ah for us. That's like our bare minimum like for for majority of our workloads. You have to have a container that runs and runs locally and also has configurations to run in given environments being a stage prod dev whatever.
00:12:44
Speaker
um And that's kind of like our our contract to developers. So for them, they see a lot of freedom. Hey, I want to write a Go app. ah You can. I would encourage you as development leads to consider how you maintain Go because we've never done it before. But majority of our stuff is dot .NET and Python or something like that, right?
00:13:03
Speaker
um So we give people like the prescribed golden paths, but if people need a business reason to go and forge their path There may be a valid reason hey this tool that we have to buy only works with go or only works with gosh like somebody PHP I was kind of kidding myself that but like in reality is it's it's really what's necessary for the business You know who are we to dictate what we can and can't do if we're building a platform. We don't care It's a means to an end And i I like that division of, okay, give us a container, you know, and and that's your way in because that kind of does open up the doors to, I really don't care what you use to populate your container. As long as it runs in that container, we can get it deployed and out on onto environments. Right. So I could, I could write it in basic and, and put it in a container. And as long as I have my, my.
00:13:56
Speaker
port in and in my outgoing port and it's able to communicate. I don't know how you would do networking on basic. That's it. You would use standard in and standard out. Oh, that's right. Yeah. Yeah, you would you run um you run the tunnel or something. and Yeah, you'd have to write like a little title like a little oh boy. ah ye Yeah. Yeah. No, but I mean, like right put we'll put, you know, flat files in an S3 bucket. Yeah. Oh gosh.
00:14:26
Speaker
I mentioned it earlier, but I mean, in reality, like, yeah, and and and I know for the folks listening, like, you just like oh my gosh, just raw containers out in the world? No, it's all part of the platform. Your security teams, your quality teams should have checkpoints in there. For example, it's not uncommon for our security team to say, hey, you're using an unsupported container or release version.
00:14:45
Speaker
Now, with that, everyone has to understand it's, you know, kind of going back to the concept of operations. You have to have supported modules and what you support. Hey, you're trying to use a version of PHP that isn't supported. Why do you need it? Get security approval. Okay. Understand we have this tool legacy. We're going to be moving to it in three months. Okay. Got it. You got to do an emergency patch. All right. Makes sense. We'll take the risk.
00:15:08
Speaker
But at some time, maybe the answer is no, you can't use that version of Go. Why can't you upgrade? Oh, um I really don't want to test. I got to get this out the door quick. No, sorry. Please, update to the version of Go that we require and you'll move forward. And oftentimes, it's nothing burger, right? Just other than like, oh, update the container image and stuff. But yeah. that it's and Interesting that, like, I don't know, we've kind of talked about this topic a couple of times on on the podcast. like the The whole notion of like, you build it, you run it, like that back in the day, that was like the whole DevOps revolution was, you know, you rebuild it, you run it. And it seemed like there was almost like this, you know, we we embraced it at first from the developer side and and that became like, that's the nirvana that you want to do it. But then you started to see kind of a retraction of that, right? The pendulum swinging in the other direction where the developer's like,
00:15:58
Speaker
and I got in development because I want to develop. So it feels like that's where these like platforms are starting to come in as it's like, okay, all of those things that you need to do to build, to run it. We're going to put that behind this platform ah or build a platform that supports all those things and lets you get back to doing what you love, which is developing software. So I don't know. It's just an interesting how that that pendulum kind of just swings back and forth. to And I think part of it is if you don't have a platform, um,
00:16:23
Speaker
Because I've, I've, I've worked with clients that don't have platforms for their operations. And you spend so much time getting your stuff to run an environment. You send like the one of the clients I was working with, I spent like 80% of my time on the operation side and like 20% of my time coding because I could get it running in container and in my local system, but to get it deployed.
00:16:54
Speaker
And the build was fine, but you know getting the deployment squared away and making sure all the the um auth systems and everything were configured properly and getting ah connected to you know the other third party services and all of that, just was without a platform in place to help you do those things, it is a large chunk of a person's time.
00:17:24
Speaker
And i think that's so if you don't have a platform that's helping you do these things, then I believe that you end up in the spot where your developers that are DevOps the developers are being forced more into operations because you don't have those tools in place to streamline that process.
Abstracting Platform Details
00:17:47
Speaker
Yeah, I mean, and think about it. like you know We're talking about the pendulum shifting. I mean, it's its really ultimately it goes down to like what motivates and what's the driver behind a team. And engineering and software development leaders really focus on velocity shipping features for shipping code. And really, anything else is like an abstraction. I mean, um I always like to refer to this like moment. It's kind of like the aha moment for me when I was kind of building platforms. Kelsey Hightower is at KubeCon talking about YAML to the to the audience. And it's like, who here likes writing YAML?
00:18:15
Speaker
And none of the developers raise their hands. They see operations people. No one cares. So abstract it away. So to me, like that is just kind of like when you as an engineering leader say like, Oh, what's going on? What's taking so long? It's like, Oh, I'm stuck making it a three bucket. It's like,
00:18:31
Speaker
Why? like oh Is the code done? No, the code's done. It's just S3's not working or something. or I don't know what settings and flags I have to put on this bucket to get to work. And that's really why that platform exists. Can we abstract away all the implementation details from our engineers? Not because they don't understand it, or they don't. They're dumb. It's just that, can we have abstract it away so they don't have to care about it? If they really want to care about it, they have a spare cycle, go ahead. But I never want to hear, oh, we're delayed because the the platform god doesn't know how to build an S3 bucket or something, right? That's why tools like Terraform and and Puppet and all that stuff kind of came about. It was kind of kind of helped solve that in my new show of like, all right, how do I get a file into somewhere? How do I create a resource? How do I keep it statically defined? um And that's why we see, I think, see the advent and rise of those tools. Yeah, I think it's, I love how you phrased it, like you established kind of those golden paths. And I think that's that's one of the big struggles, right? when you're When you're trying to develop these platforms for people is,
00:19:31
Speaker
you know and And I've been on those teams right where you're you're building the the common infrastructure and the common tools that other people are going to use. You're like, OK, this is this is the the thing that everybody should do. And it's very smooth and easy. And everything works great together. And then one team is like,
00:19:44
Speaker
Yeah, but we need to do this. So it's it's tough, right? But it's, it's interesting that the way you phrase it, like that golden path, that seems to be like the way people approach this mostly is like, make the right thing to do easy to do. And that would, that will, you know, we're developers are lazy. We're going to do the easy thing. So it's more likely you'll get an adoption if you just make it very, very easy to use.
00:20:06
Speaker
Yeah, and and to build on that, is it's all about um automating the majority of what you do. Like, if you think about most places you've worked at, or even I've worked at, look, I've worked in Java shops, I've worked in Python shops, I've worked in in Go and all all these shops. Focus on what the core deliverable and what the business does 90% of the time. Most of the time, a developer gets tasked with a new microservice or new service in general. What are they writing in? Oh, we're a Go shop. All right, I'm going to make a new Go container, deploy the Go container Kubernetes.
00:20:36
Speaker
figure out how to automate what 90, 80, 70, 60% of your workflow is, so then you can focus on the hard stuff, but the 40, 30, 20%, right? Because once you start focusing on that, you may actually see patterns that you didn't even think about. Oh, wow, the rest of the 20% is liquid base or database changes. Why don't we make a liquid base pipeline? All right, cool, that's another 10% of that. um And then you can start seeing patterns kind of through the weeds. get ah Get the noise out of the way so you can start focusing on those core principles that don't have maybe automations in place.
00:21:05
Speaker
Yeah, yeah he' definitely never suggest starting with the special snowflake, right? it's like it it's It's a recipe for disaster because you're going to be in in the weeds so much dealing with these like one-off situations that you're not going to get your You're a major part of the bell curve, right? You're going to miss your main audience. yeah We're all special snowflakes though, Aaron. All of us, every team is a snowflake. I'm not a special snowflake. I'm too salty. Well, I also build on.
00:21:41
Speaker
Don't even, even if you're not focusing on the special snowflake and focusing on the general, don't pick the largest one either, because that's also the most visible yeah for successes and failures, higher risk and reward. I mean, a lot of places I've been, it's very tempting, like, oh, I'm going to start automating and DevOpsing, quote unquote, the, ah the, the payments API. This is going to be huge. I'm going to get so much recognition. The team is going to get big get a big win. And then the first pipeline failure happens and someone says,
00:22:07
Speaker
Why are we building a platform again? Like what are you guys doing? Like what's the outage? Uh, now if you started with like that minor brand website and that caused an outage for like a few minutes, you're like, Oh yeah, it was down for five minutes. Okay, cool. Yeah. Kind of gets a little bit of a different perspective, right? Well, and managers tend to like to push those, those larger projects too, because they see the dollar signs of, Oh, if we automate this, this is where we can get the most bang for our buck.
00:22:32
Speaker
right But it's like, no, if you slow roll it, you'll yeah it'll work better when you get it to your big project. So you don't end up in current costs and in you know putting out fires. Speaking of you know platforms and everything, you know what a frog's favorite platform is? Reddit, Reddit. Oh, boy. Wow.
00:23:02
Speaker
When you're thinking about these these these platforms, like yeah obviously you want to give them a lot of you know them being developers, the tools they need to get their jobs done, but how do you constrain that so that they they don't have so much freedom that they can break the world?
00:23:21
Speaker
Yeah, I mean, we just give all our developers root access stuff that we're not kidding. I actually had an operations person say, I'm going to make it so you can't do any work at all. And I'm never going to touch your computer. I'm like, what are you talking about? I'm going to give you root password. I'm like, oh, no, you're not. No, you ain't. Oh, no, you ain't.
00:23:44
Speaker
<unk> When you think about it as a whole, I mean, the real answer is like, yeah, look, it's a lot of risks. um I mean, personally speaking, it it starts with a lot of education and training um and and kind of understanding as to what the platform can and can't do.
00:24:00
Speaker
Out of the box, you know the way we see it is we build a platform with empathy for our developers. If a developer can cause a systematic break, um it's the fault of the platform, not the developer. um right um I know that puts a lot of burden on those people building those platforms and those delivery teams, but that being said, like there should be some level of rigor or at least understanding of what they are going to change with that change.
00:24:23
Speaker
right So um I'll use a kind of a tangible example of like identifying IAM permissions. Maybe there's a baseline of IAM permissions that you can associate with a deployment. like Let's say um I am working on a very small segment of a database and I know that my team always needs read access to it.
00:24:42
Speaker
Great. Maybe that's a permission that could be enabled to grant and maybe even reviewed by a security team at the start. Maybe it it becomes a standard change in the future. But all of a sudden you start granting like right access to a component and you start modifying things. Now that's his perspective. Like shouldn't there but have been a check and balance? Should there maybe have been something to say, Hey, can a database administrator review the proposal to move from read only to read and write?
00:25:07
Speaker
um do we know the impact can happen um and i think that's kind of where you know you have to think about what what your business lines do cuz everyone's a little bit different some people are in no sequel some people are in google some people are in a w s really yet start thinking about. What are the.
00:25:22
Speaker
expected things that my platform should be able to allow you to do? And what are those kind of one-off exceptions? And then eventually, if those one-off exceptions become standard enough and there's enough maturity, you kind of start thinking about scorecards and giving people that that freedom of, all right, cool. Hey, you've done this multiple times. You've done the right thing. How do we make this in a safe way? Again, not saying just give full on access to people, but in a safe way to continue to to operate and again, drive that velocity because it goes back to that velocity. Engineers care about one thing, shipping features, getting things out there.
00:25:52
Speaker
Anything in that process that slows them down is just seen as a bump in the road. Yep. And it makes it harder to estimate too, honestly, because a lot of the, the estimates of work that comes in falls to the development team to say, okay, yeah, we can get this done in this amount of time. And when you have that giant question mark of, I don't know how long it's going to take to deploy. Um, I'm going to have to pad my estimate for that because it should work.
00:26:22
Speaker
should and you have to air quote it. um But and you get burned a couple times, and then your estimates start going up and up because because you have that and fear and uncertainty of actually deploying out into an environment that should be cookie cutter. in And it's like, we're not doing anything special, but I know I'm going to have to spend you know three or four weeks getting that initial deployment done so that it's stable enough that our CICD will be able to deploy to it and and not have issues on a regular basis.
00:26:55
Speaker
And so that's it's definitely something that i I've done it both ways now. And I'm like, I so miss having a platform in place that that takes care of that kind of stuff for me, that doesn't allow me to um go in and screw things up, but also gives me the tools enabled to unlock jam stuff as well. Because a lot of times too, its you know especially when you start talking about cloud environments, there's so many things that can kind of make your deployments not successful, especially if you're doing it repetitively. An example, we were using API Gateway.
00:27:43
Speaker
for that. And um we ran out of VPC links on a continual basis. So we had to go in and then scrub all of our b VPC links. And you can't do that if they're tied to a load balancer and all this stuff. So you'd have to go in and spend, you know, a day or two just going in, finding the load balancers.
00:28:03
Speaker
ripping them out, then clearing your VPC links in and then redeploying. And of course, you know, deployments take, you know, 20 to 30 minutes at a time. And so if it doesn't work, then you it's just like.
00:28:18
Speaker
ah ah Yeah, I mean, you mentioned, you know, especially like, it's actually, that's actually a sign of maturity when like, whenever I'm implementing a platform in a new organization, like, that's actually one of the signs that I look for. um You know, typically when I arrive, everyone's talking about deployment windows, and I say, I arriving is like kind of joining a new organization. Everyone's like, Oh, yeah, Tuesday and night at nine to 10 o'clock, we're going to be doing this change.
00:28:47
Speaker
um when you know announcing that's fine but when people stop debating the time and how long and people start talking about the impact and the change itself more so than the time the physical time um that's usually a sign of maturity for me when people just start talking about like hey we're rolling out the new patient check-in experience and that's going to deliver so much value. that at the thought When you start talking about that versus like, oh, how long? Duh, it's a platform. It's going to take five minutes. ah you know It's like, any questions on it? Oh, no, no. Okay, cool. Next. you know that's That's usually a sign of like maturity in my view. Yeah. and and I've worked in those environments before and it is it's it's so nice to have the right tools in place for that. it just It's a game changer. Yeah.
00:29:34
Speaker
You know, it's because working without it is, it's kind of like, it's, it's like going to the dentist. You just dried it. and yeah Um, well, which speaking of dentistry, the, one of the, uh, the major things that the aspect group does is dentistry. Isn't it?
00:29:54
Speaker
Yeah, ah we have Aspen Dental as kind of one of our larger brands under the parent umbrella. um So if you are a customer, thank you so much for visiting and always appreciate your business. Don't forget your candy. Hopefully make you smile a little bit, not as yet scary as Dental's. Don't forget the candy on you know Valentine's Day is coming up and Easter and all that. Yeah, gets get lots of candy. and we We'll check out for cavities any day. That's right. yeah it's like yeah i would
00:30:24
Speaker
You know, you always know if you've got like the ambulance chase or dentist in your neighborhood because they're handing out things like, um, frozen Heath bars and um in giving the kids like hockey pucks. Hockey players are a dentist best friend. That's a good idea. Yeah. youre guy That's a, that's a good one.
00:30:46
Speaker
It's, it's like the guy that goes in and always make sure he writes one bug for every 10 stories. nice Job security. Yeah. So we we didn't really talk about it much, Joel, but like how, what, what is your career path been like? how How did you get to doing this thing that you're doing now? and And even like at the Aspen group, like how, what would that look like? What was your journey like?
00:31:10
Speaker
Yeah, I mean, honestly, I'll just start really early in the kind of brief, but like, you know, I started off as like a systems engineer and then became an applications engineer early as like my career. And in both places, I kind of just recognize them like.
00:31:24
Speaker
There's just got to be a better way. So we we did things like puppet, we did things like chef, we did Ansible, right? um We automated application rollouts with that stuff, but it still was so manual. Like it felt like there are always so some cog in the wheel. And the more important thing was it still felt like I was just getting poop thrown at me by developers. So it's like, there has to be a better way where we kind of work together and bridge that gap.
00:31:48
Speaker
So at my time at GoGo, we built a lot of these like automation platforms. We used a tool called Spinnaker from at the time, Netflix, and now it's an open source tool. um But now the industry is all Kubernetes. So we're looking at things like Argo CD, and that's kind of our focus right now at at the tag group. We're moving ah to Argo CD, looking at how to continue that platform experience. Again, developers shouldn't care whether we're using Argo, Jenkins, GitLab, or Azure, DevOps, whatever.
00:32:15
Speaker
As long as the platform kind of builds that experience where they can just commit code and focus on ultimately what they're delivering and building. That's really what the experience we're trying to build out here. How do you like Argos so far?
00:32:27
Speaker
um It's good. it's ah it's It's come a long way since I've had it. um you know For me, the big thing for me is shifting from like a tool like Spinnaker, which focuses very much on pipeline delivery and software rollouts in a pipeline, to GitOps. It is a little bit of a shift, but there's a lot of awesome things kind of with Argo events, Argo rollouts. A lot of the stuff that I had an issue with Argo early on in the project are getting resolved. and Now, I think that's why you're seeing such a mass adoption. I don't know about you all,
00:32:56
Speaker
It seems like everyone's building either an extension or building on Argo, which is great to see. And the beauty of it is the CNCF and all the Kubernetes community is basically behind it right now. So, so far, we're really liking it. We're liking what we see, um keeping things in sync with a repo. It's almost like, think of it as like a continuous terraform apply running against your cluster to make sure that things are the way you described that they should be, um which is, as an operations person, awesome.
00:33:23
Speaker
Yeah, I really liked that it. How about, you know, and I think it's interesting that for early on when it comes to those like more pull-based, you know, here's yeah declarative, more pull-based model, I was using flux and they did like flux. I thought flux was really cool, but yeah, Argo, it seems like it's just, it's kind of one out. Is that kind of the, where everybody's at in the CNCF cloud native? I mean, I wouldn't say I, again, like being on a podcast and recognizing I'm saying recommendations, like I would say like never, no one should ever be ashamed as to what they do.
00:33:53
Speaker
and pick, but I think like Argo right now, to me, is kind of the popular favorite. But i mean at the same time, like you know again, it goes back to some some early meetups I've had when we were talking and people were describing the perfect architecture. Don't just look at what other people have. Look at what you have and how to enhance it. And if there is a need to move to things like Argo or to move the things off to Flux or whatever, pick the tool that works for y'all. Because at the end of the day, that's the biggest thing, operational support. Because and you don't want to be stuck in kind of the situation where you're I don't know, kind of between tools and you don't know how to maintain it, you know what I mean? Yeah, the one thing I liked about Flux and it it was a native feature of Flux that's not in Argo, at least not yet. it's I think it's like a lab or something like that is the the whole notion of like when I deploy a new version of my container image, it scans the container image repository. And if it sees there's a change, it'll go change my manifest for me in Flux. It would do that automatically, but Argo doesn't have that, at least not.
00:34:49
Speaker
the version I was using when we installed it. um Yeah. And I forgot what it's called. They announced there is something in kind of the works sort to do that now. I forgot that the event notifier. it's ah it's It's definitely, there was a KubeCon talk this year. I remember seeing it, and but definitely check it out if you haven't.
00:35:05
Speaker
and Okay. Yeah. That was a really cool feature. I liked that. I thought it was really nice. Cause right now we have to kind of like script it. I mean, we were going to use them like said or knock or something in our scripts. Yeah. We are to to update our manifest, but, but I do like our go now. Do you use the UI?
00:35:20
Speaker
Uh, yeah, we still use the UI right now. So to kind of keep things in sync, it's, it's really helpful. I mean, I wouldn't say you should do all your operations in there. You should definitely use things like, you know, your, are your Git repos and stuff like that. But I think the nice thing about it is the visual tier. Um, look, I, I'm going to be completely honest. Kubernetes is new to the Aspen group. Um, I have engineers learning Kubernetes for the first time. We're not perfect.
00:35:41
Speaker
um Again, I'm never going to fault someone for not knowing the kubectl command exactly and not knowing that to do dash all namespaces and stuff, right? I would much rather people learn, grow, and then figure out, oh, I don't need the UI anymore. I can take the training wheels off. Give me the command line, you know? Yeah, I like the the visualization that it gives you too of like your entire deployment of everything that's out there, how it's linked together and all that. And I think it's really nice.
00:36:05
Speaker
I think United Airlines or American Airlines, one of them ah built actually a pretty cool, it was a KubeCon prior. They were actually talking about building components inside of Argo's UI. um It was actually pretty nifty. So think about like extending things like after you deploy, you could also get like a Grafana dashboard for lack of a better explanation along with it. It was just kind of like an external dependency being mapped inside of your Argo UI, um which I think was again, pretty nifty to see. That is cool.
Personal Book Collections and Life Journey
00:36:33
Speaker
Yeah, i I love a good UI around that kind of stuff because the CTL commands are are are are different for everything. and and its just
00:36:44
Speaker
there's There's definitely a learning curve, especially if you have to work in mixed environments where you know you've got Docker for some stuff and cubes for another. and it just it's it's ah It can get confusing. It's like, okay, was that a GitHub command or was that a cube command? was that a socker commit How do I how do we do that the that Docker command again? and It's just like you almost have to have like a a playbook of ah different things. I think there used to be a ah book called like, ah and maybe it was an O'Reilly series. So they're like playbooks or recipes. They call them recipes. Yeah, recipes and cookbooks. Yeah, yes recipes and cookbooks. That's right.
00:37:27
Speaker
um And those were the best because you're like, Oh, what was it back when we actually used, you know, paper books.
00:37:38
Speaker
I still like it. I do too. before I love getting a nice paper book sitting down on the couch. I don't know. Like the Kindle I know is lighter, but I don't know. I just call me an addict for the way this page sounds or smells when you turn it. It's like, it's great. yeah It's it's the the tactile feel of the page. It just, there's something comfort comforting about it.
00:37:59
Speaker
um I'm kind of the same way as it's like, I have i in my bookcase behind me, I have, you know, all kinds of, of different books. I just like, am I, am I going to use them? Maybe, maybe not. But I like having them. It's, it's a comfort thing. You know, honestly, to me, someone described it as almost like a, a, a, a encyclopedia of your life. Cause when you look at your shelf, uh, you can probably see like, I remember when I opened that book. Wow. That was 20 years ago.
00:38:28
Speaker
Oh yeah. Or I remember that book that was 10 years ago and you kind of start building like a portfolio of how you got to where you're at. So um I like it for that. At least to me when I pull up a Kindle, I'm like, I'm not just swiping through like a thousand books. I'm looking at like my life basically, i have things I've used and learned over time. That's cool. It's a cool way to think about it. the yeah The stories you read tell the story of you. That's cool.
00:38:54
Speaker
So the one thing and I know we mentioned, we touched on it a little bit, but like, and one of the things that's really cool about this podcast is is we interview people from like all over the place. I mentioned different domains and stuff. You mentioned dentistry. I know you all have some other brands, but like, are there any, is there anything like that you would consider to be very unique about what you do because of the domain you work in that the might the interesting challenges that you have to face that anything like that.
Managing Brands Under TAG
00:39:19
Speaker
Yeah, I mean the cool thing about like TAG right now is that we're basically kind of overseeing multiple brands and those brands are unique. that We don't just have a ton of dentistry brands. I mean we do have things tangentially aligned to it like clear choice for dentures and dental implants and stuff like that. But we also have brands such as like ah chapter for beauticians and like you know things like Botox and stuff like that.
00:39:42
Speaker
We also have a brand for dogs and vets. you know We have a called Love It, so um basically a new vet brand that we're trying to expand. And ultimately, we have things like Urgent Care. um you know Hey, ah yeah ERs are getting expensive nowadays, and sometimes they can't even see you on time. If you just need something quick, check it out. We're there for you, too. So it's actually pretty cool to see. it That being said, as you can imagine, we just talked about like five different industries in in one. um The context switching is um It's not only challenging, but it's also rewarding. um You're kind of working on multiple business lines, but you're also not rebuilding a platform for everybody. you're you If you do it right, you build a podcast platform that can support everybody. And that for to me and my team, I think we really value the opportunity to have a chance to do that. It's kind of cool. It's kind of fun.
00:40:30
Speaker
Yeah, we had a guest on recently, Erica from the the Yum! Brands. um you know they They have a similar idea. of like They have multiple brands under one roof. And she she came at it from more of a user experience perspective, from ah from a visual perspective for the users. But you're now you're coming in from like a platform experience. But she had the same sort of experience. There was a lot of things that are across you know brands. There's things that are unique, of course. They have to be. but You are able to get a lot of bang for the buck and and, you know, commonality among all the different brands and be able to support them. And you don't have to, you know, fork everything seven different times or whatever. And health aside, I mean, it' lessons learned. I mean, like like I mentioned earlier, Aspen Dental being our largest brand.
00:41:12
Speaker
There's a lot of lessons learned. We have hundreds of millions of patients, right? um Versus like something, maybe a smaller brand. Love, it's a brand new brand. We're starting to grow our footprint. It was formerly called AZ Pet Vet. It was only in Arizona. Now we're starting to talk about national scale potentially. um the the The lessons that we learned trying to scale one business line is almost the same. I mean, not exactly the same, again, very unique, but a website's a website, a login is a login, a patient's a patient, if you mean even though they may have fur and four legs.
00:41:40
Speaker
Uh, they are, you know, it's very much a patient. We respect our patients and we want to make sure we do the same things, even if it's, you know, uh, data security for a dog. I mean, I think we would probably all not care if our dog's gender was leaked on the internet or something like that. But at the end of the day, it's, we want to keep that same level of rigor regardless. Right. It's, it's with personal data that we want to make sure we protect our customers. I mean, if a dog's personal information gets out there, it's rough.
00:42:07
Speaker
Wow. oh no Well done. Well done. Wow. okay Just let that soak in everybody. yeah yeah Let that one just marinate on that one for a little bit. Well done. and so and You know, the with The idea of a platform is exactly that, though, is ah it's it's something that's supposed to be more generic that fits a lot of different um inputs, right? It's supposed to be kind of the the template for how you're going to do things. and because we've always we've always heard you know Because there was this pendulum sling back when we were doing software platforms.
00:42:56
Speaker
that you know it's like, oh, well we don't want to create a platform for everything. Well, yeah, right, because then it's not a platform. It's just another application. right So you want platforms for your your major groupings of things like brands or domains or something like that. So if like if I'm in this domain, I should ah should be able to use a platform for that and maybe take that as a starting point for a platform for another domain.
00:43:26
Speaker
and and reuse a lot of that. And I didn't even argue that there's even like this concept of platforms for like the hardware and software components. I mean, we talked about it earlier with the TVs, or sorry, this the fridges and stuff. Do we really think Samsung just rolls out updates and doesn't have it a monitor as to how that updates going? Or are they just going to brick a million fridges and just refund everybody, right? there's that's definitely i mean I mean, I guess maybe I should be careful. I didn't look at recalls lately, but ah at the same time, I'm sure there's there's some level of rigor as to how that hardware is updated and monitored. i mean
00:43:58
Speaker
Even in our industry, like, you know, the other day we we're talking about the air compressors and stuff for our drills for in the, in the dentist office, like, Oh, what are you going to monitor the air? Yeah, you can technically, uh, I guess the vendor actually sells a monitoring component for the air and the compressors and all that stuff that we can remotely monitor all our offices. And we're like,
00:44:15
Speaker
Oh, uh, I'm sorry. Uh, let's, let's, let's do that. So like, we're always trying to like make things better, even if it's hardware, right? Uh, that experience. And again, different audience, that's completely different. So you're talking physical installations and stuff, but to even get that visibility and that level of into like a common platform, you start thinking about things differently. You could actually use that as a marketing tool too, is like,
00:44:39
Speaker
Um, you probably don't want to go to this, this dentist. He's drill happy. like Much more air going on over there. We noticed that drill being used a lot out of this dentist office because they're so good and they're busted. For some reason, it's always towards the end of the year when you're, you know, you're deductible is about to wear out. You've already spent your deductible. Interesting how that works.
00:45:09
Speaker
um and and So that's like kind of a predictive maintenance thing, right? So you would be able to monitor the location. Yeah, I mean, it's it's it's it's no different. I mean, at the end of the day, it's about being proactive. I mean, ah how I mean, how do I work this like, you know, we all like to bash our ISPs, right? And so like a house and much of things. But has you have anyone like really opened an ISP ticket lately? It's it's pretty impressive, even the electrical grid, like,
00:45:32
Speaker
I've gotten outages from ComEd before I knew that there was an outage that hit me, and out as our or ComEd being the central Chicago in an um electrical company. And when you can go to like a website and see a status map and they actually have like a marker as to what areas are and impacted and you're like, this is awesome like I mean I'm so upset that I can't like cook my food or something but at the same time I'm like wow I actually know what's going on I'm not just calling 1-800 numbers 24-7 asking for a status update um at least as a customer I feel like there's a little bit more rigor and and and they care about me and I think that's the same thing we try to do if what we both our platforms and observability strategy we want people that are dentists our customers to feel like hey we do care about you we recognize things break all the time but it still sucks that you're impacted by this
00:46:21
Speaker
And for those playing along at home, ISP stands for Internet Service Provider,
00:46:29
Speaker
which I actually have, I almost submitted a ticket just ah a few weeks ago. They went through and in they kind of bricked our internet for a little while. um And that was, ah it was a problem.
00:46:43
Speaker
Uh, they were doing, but they were doing hardware upgrades. And so it was, it's kind of one of those things is like, well, if they're coming through and updating stuff, it's a little hard to have uptime. Never mess up a tech guy's internet. We know the ping. and la Well, especially when you work remote, right? I was like, well, there's my lifeline. I was in working off a hotspot. Just. Ooh.
00:47:04
Speaker
It's rough. I could do it, but don't expect video on my on my calls, right? So the the when you have a setup like you guys do where you have like ah multiple satellite offices and all of that, primarily your locations are all over the place, right? so That, you know, that's a different level of support. You kind of have to, it can't be like the dentist office. We have a server go out or we have an outage in the data center that now that dentist office can't take care of patients. You have to kind of have be in that offline mode, so to speak. Do you guys, is that a challenge for you guys? Yeah. I mean, look, the company's now 26, I think 26 years old at this year. Like we, they've the the company's been around for a long time, right? You know, uh,
00:47:49
Speaker
There's definitely a lot of things that they've learned over the years and there's a lot of fallback processes. I mean, think about 26 years ago, none of this platform engineering knew how existed, right? So there were papers and faxes and stuff like that. So there's always like a way of keeping the lights on. um That being said, the the the burden always comes in and reconci reconciling that. Oh, now we have to process all these faxes that were done manually or these these ah these signed paper documents, right?
00:48:14
Speaker
We never want to be in that position because that's that's burden, that's time. I mean, I would hate to be the person in that situation. But I think as we kind of go, we start thinking about, all right, well, what's what's the level of severity that we're at? And I think establishing those principles are are big. I mean, even outside of health care, I mean, I encourage that everyone think about that. Like, you know, everything is going to be a P1. I get it. You know, everyone thinks that everything's an issue, a number one issue. But really, as you start maturing, you start thinking about, well, how How much do I have to have on this? Do I need everyone in the company handling this issue right now? Can I have one senior engineer working on this? Do I need to hire an external vendor, like ah like a power expert, to go out to that office? And that's like something that's out of our control. We don't have expertise in-house. it's yeah Someone crashed into a light pole. We've got to go call a 1-800 number to someone to fix it. um And that's some of the challenges. There are things that are just outside of your control. um you know Hey, someone crashed into a ah power converter, and now your office is offline.
00:49:12
Speaker
Hey, guess what? You're out, but also Starbucks and and the the old tub beauty is next year is out too, right? Or something like that. So you guys are thinking about stuff that too. We were, we were brainstorming that kind of, um, the, you know, catastrophic event recovery type of conversation. And, and we had some sales folks in the room too, because, you know, they've, they've, they get calls from, from customers out in the field. And one of them was like,
00:49:41
Speaker
It'd be great if we could get notified if the power goes out. I'm like, um what's going to send that email?
00:49:58
Speaker
yeah And they're like, oh, yeah. but Yeah. But you know it's it's that kind of thing. it's like Sometimes you just there's there's not a software solution for um what it is that the that you're dealing with. Sometimes it has to be boots on the ground.
00:50:18
Speaker
Yeah. Or sometimes you have to develop it. Right. Like, I mean, right in certain cases, like, all right. Well, we, we know the office is going to be offline. So their phoning home is not going to be possible. But do I have a monitor from home to, to, to, to, to phone too. Right. Like, so if all of a sudden I start seeing, Hey, the entire East coast offline. Oh yeah. There's a major hurricane kind of ramping up the kind of coastline. That's what's going on. So to be expected. Mm-hmm.
00:50:48
Speaker
Not like the, the East coast power grid failure. It's like, yeah, well, we're not going to be expecting any, anything working over there for a little bit but of a dark spot. Let's just say that not many people were going to get their teeth cleaned during a major hurricane.
00:51:03
Speaker
yeah Well, I've got the windows all boarded up. So now it's time to go get a crown. yeah <unk>s Perfect time for that. Right. Yeah.
00:51:18
Speaker
All right. Well, I think it's time for our next section, James. shebo skin but ships give everybody we got to tell if shebo was good yes Ship it or skip it. So ship it or skip it works like this. We're going to give a throw out a topic and we'll go around the room. Every, everybody will give their idea, their, their opinion on whether it should be a ship it, which means it's ah a good idea. Yes. Keep this around.
00:51:48
Speaker
or skip it. Now we don't need to be doing that kind of stuff or we need to get rid of it or phase it out, whatever. So the first one is build your own platform. Oh, let's start.
00:52:05
Speaker
No, you can start. um This is, ah I think you should ship it. um You should ship it. Well, let me rephrase this. I think you should skip it because there's a lot of established solutions out there that are coming about. And especially if you don't know what it takes to build a platform, you definitely want to have some expertise of people that are supporting other companies to do it. That being said, if you have a significant staff that already can do it,
00:52:32
Speaker
um you should probably build it yourself, because you're going to know your business lines. That being said, the happy path of finding the the platforms. you know I've seen many companies like Port or something like that out there, which support many different data models. um So you can build that custom experience to kind of get a custom um a platform. um So I think it's it's just a little it's a little bit of both. um So I hate to kind of give it a 50-50 on this one. But I think you have to know what your what your what platform, what you're building. I mean, there isn't a vendor in the world that understands your business better than you.
00:53:07
Speaker
I'm going to be ah more towards the skip it with the with the idea. It's similar vein of thought. It's like you want to get you want to leverage what you can get off the shelf to the the best ability, but most of the stuff is customizable to the point where um if you can find a solution for your platform that allows you the freedom to integrate it and customize it the way you need to. I think that's that's going to be better but in for the same reasons that you you do need to have a team that knows your domain in there so that they can integrate it and customize it properly um for you. So there's still going to be a bit of a build in there, but I think that square one would be the the purchase something and then mold it into your domain.
00:54:03
Speaker
Yeah, um some of this is for me, it's like a ah temporal answer. So like if you asked me five years ago, I'd say absolutely ship it, build your own. There's not enough mature products out there that that do this the way you would need it. um But now I think maybe the opposite is true. But I would also maybe put another axis in there and say like it depends on the maturity of what you're building. If you're if you're really dealing with some some gnarly, nasty legacy stuff, that's there they all it is is snowflakes.
00:54:33
Speaker
A platform that's that's built for more like cloud native solutions from the get-go probably isn't going to fit like a glove for you. You're probably going to need to go custom until you get yourself more migrated to be more cloud native and you've you've lifted and shifted or or you know rewritten some of those things. And then those platforms will be a little more of a good fit. I don't know. that's'm I'm a little half and a half, but I guess myself. Yeah. ah you mean You bring up a good point. Legacy systems are generally snowflakes because they didn't have a platform to work off of to begin with. um So yeah, i I would put the the footnote on there, that you not not intended for legacy systems, so right?
00:55:18
Speaker
or writing it yourself is not intended for mature audiences, right?
00:55:27
Speaker
It's kind of the opposite of the, uh, like that. Um, all right. Well, oh, our next one that, that, uh, we were talking about the term cube cuddle instead of cube CTL.
00:55:46
Speaker
um There was actually a round of this at KubeCon. I mean, look, for me, I think, um ah ah I skip it. um I say KubeCTL, KubeControl, all the stuff too. I mean, look, I'm a analytics guy, I'm a systems admin guy. SystemCTL, system control, right? Think about the the old Unix days, right? That's that's really what it is. KubeCuddle is cute, I love it, but not, I'm skipping that, sorry.
00:56:17
Speaker
Aaron's shaking his head. You guys can't see it, but I can hear it. yeah with i'm I'm totally a skip it on this one. i'm I'm sorry. If I'm telling, if I'm talking to somebody outside of the tech domain going, Oh, what do you, you've been working on. I've been working on cube cuddle. They're going to think, ive ah you know, like I need sensitivity training or something like that. It's just like. No, I'm not, I'm not throwing that in my vocabulary. I'm sorry. It's hard enough to get people to understand what it is I do for a living. I don't need to add something confusing like cube cuddle. It's like, Oh, yeah.
00:56:57
Speaker
buts right we're going to keep an eye on this guy end up on a list somewhere no skip it absolutely skip a guy over there with the care bears t-shirt on in the neck beard and that's that's a little weird um i i've I've used cube cuddle, that's just what I say. I don't know that I'm necessarily for it. It it seems a little absurd to to use, to say cuddle, you know, because we we think of what we do is very ah very hard at like a,
00:57:33
Speaker
a hard thing. It's manling yeah i not difficult but it's like yeah so manually that we that we sit at a keyboard and, you know, peck away at these little these keys. It's tough to depress them. that I have counts on all my fingertips. I type with aggression.
00:57:50
Speaker
now now we' Very interesting to see is your viewership after this, like, oh, these guys, these, oh, that's it. I'm out of here. Every at the end of the day have fun with it like I get it like I keep CTL keep cuddle I have I've been guilty of saying it too I think like for me It's just always been keep CTL or keep control like I mean that's that's that's a little bit of a stretch but keep CTL for me Well, I am the beautiful co-host, so it would make sense. Yeah, everybody wants to keep cuddle, James. Yes. That's why we don't go video with our podcast. We're staying audio only. This sounds like HR points in the cubicle. is yeah
00:58:31
Speaker
but so let's I can't do it. I can't do it.
00:58:39
Speaker
All right, what about this? This one's an interesting one for the for the um community, for the the Kubernetes world, Kubernetes, however you say it. um Cloud Foundry, ship it or skip it. Oh, I her um I Never used it ah so I feel like I shouldn't comment on it ah So I'm gonna say skip it for now with the unintended bias. I have never used it um But that being said, I'm sure someone has used it. So who has used it? Well James and I both used it at the same place Yeah and
00:59:29
Speaker
You know, and I was there from the the homegrown solution and then also the transition after Cloud Foundry. um
00:59:45
Speaker
I think it really depends on where you are on that maturity curve, like we were talking about before. If you're if you're on that maturity curve of where you can build it yourself um and you're more on that end of the of the spectrum. I think you could you probably don't need something like Cloud Foundry in place. um I didn't mind it though. So I think you know i'm kind of i'm I'm really on the fence on that one um because I really liked our homegrown solution. I thought it was very easy and intuitive. The Cloud Foundry stuff was a little bit more
01:00:32
Speaker
So, so. Yeah, I think. Yeah. Yeah. I liked Cloud Foundry a lot. And I think it definitely influenced some changes in the Kubernetes world to a degree. One of the things I miss, and there are some efforts out there. I forget what the standards are off the top of my head or the packages or whatever they're called.
01:00:55
Speaker
Uh, but like one thing about cloud foundry was like, if you, if you want to connect to a database, you would stand up a, you know, your, your database connector, and then it would inject the connection information into your environment in a standardized way. So if I want to you know start up my Java application or spring application, it would understand and know that standard metadata or that standard connection data format. So it it would pull it from the environment automatically. Whereas now you kind of have to roll that, okay, I'm going to do DB underscore pass and like, you know, make it up as I go along.
01:01:25
Speaker
That was kind of cool that standardization was really neat and it wasn't just for databases for kind of all of the things you would need to connect to there was kind of a standard for how that connection info would get injected into your, your environment. I thought that was kind of cool. The other thing that came out of that was like, and I don't know that it was, it was Heroku and cloud foundry, kind of the, the build packs thing that was that came out of those, those platforms as a service kind of things. I thought that was cool. And I, I love how we, you know, I'm a Java guy. I love how we build our, our apps now with our containers using build packs. I think that's a really, really nice and streamlined way to go about that.
01:02:02
Speaker
I will say that I did love working with the pivotal team. They were, they were, um, they were a great team to work with. Um, so I think like from what I, from what I've, you know, from a Lucy gather, again, me not having ever used it. So sorry folks who love cloud foundry. I think it's basically gives you like a prescribed way of making like, essentially like you you said, it like Heroku app or like a 12 factor app.
01:02:27
Speaker
you know It tells you how to do it. So I'm not going to discredit it because I know everyone's on a different maturity cycle. But i i the way I see it is like almost with Kubernetes, you're expected to build that. With Cloud Foundry, it gives you a prescribed way of doing it. And but to be honest, both ways are actually fine. I mean, some places like a US small shop, do it the pretty automated way. You don't have time to, you only a one person, two person team.
01:02:53
Speaker
you don't have time for it. But if you have like a large team like you know um that that can maintain and knows the maturity cycle, and also don't be fooled by the lack of maturity you need to run these systems. I mean, Kubernetes, it isn't just you know one engineer knowing kubectl and knowing how to deploy things manually. right there's There's a whole level of maturity that comes out of it.
01:03:13
Speaker
um On that note, a 12-factor actually at KubeCon this year, they announced that 12-factors methodology is now open source. um I'm actually a part of the Discord where they're actually like starting to add things like that are more modern, so I think one of the big ones lately that's been ah coming up has been um identity into the 12-factor stack. That's a missing component.
01:03:33
Speaker
or config management, um like you know remote config management, like how to add concepts to that. So just giving that a plug in case anyone's interested in 12 factor, it is getting some some love in 2025.
01:03:46
Speaker
So is it going to be 14 factor or are they doing like brick in and break out? Let's throw out one of the old ones and put in a new one. That's a good question. I don't know what happens to 12 factor after 12 factor. So yeah they add more stuff, more factors. Make it prime, make it 13. Yeah. 13 factor. All right. I think, I think that's good for ship it or skip it. So now we'll go into ah our final segment of the podcast. The infamous lightning round.
01:04:35
Speaker
ah So the way the lightning round works is James and I are going to rapid fire some very serious hard hitting topic of questions at you ah and we will to do this in an alternating fashion. Back and forth, it'll be 10 questions. Hopefully you can make it to the end. um So far, I think we only had one person that passed out. No, no, maybe not. Was that a stroke or what did they pass out? Maybe it was a stroke. I don't know. It just hurt. yeah Like a flash of lightning, they were gone. That's right.
01:05:19
Speaker
No, in all seriousness, these are a bunch of rapid fire questions. Please keep your answers short. No explanation is needed unless provoked. Sometimes we'll be like, what? yeah Are these like, yes, no questions? or No, these are these are short answer or single word question ah responses.
01:05:43
Speaker
All right. So James, you want to kick us off with the lightning round? Yes, I can be you know counted on to do the real hard hitting questions like, how do you feel about cranberries? I like them. They're pretty good. I think they're just they have just the right amount of tart. The one thing I don't like is you have to taste. We would have also accepted no.
01:06:11
Speaker
answer yeah either way either way what's the type of triangle with two equal sides called oh snap thinking about geomery ah
01:06:29
Speaker
so not precisely no so it awfully We would also accept equilateral because technically two sides are equal. Yes. that would like not um Math it very yes ah If you were given, no that's probably not a good question. ah Have you ever tasted soap? On purpose?
01:06:54
Speaker
at all uh accidentally yeah i don't think i go around eating soap bars i was unfortunately not on the tide pod craze oh that's the best answer to that one i've heard yet nope i didn't do the tide pod craze uh would you eat a day-old taquito from 7-eleven though uh when what did i of course sometimes the best is the perfect amount of crunch you know nothing like making a taste or regret after that taste regret i love it
01:07:28
Speaker
most embarrassing store you might be seen shopping at you know what yeah like Ulta I like to like I sometimes have to buy like clippers for my hair and stuff so it's like a beauty like makeup and stuff it's like oh what are you doing buying like it's like I'm buying like clippers and stuff or Sally's beauty so it's it's like it's like a beauty supply start Oh, snow sp that's a good one. Yeah, I look at that place going, hmm, yeah, that's probably not a place I want to be caught in. I mean, I do have hair. I'm actually growing it out right now to see if I am balding or not. But fortunately, I do have some hair left. Oh, that ship sailed for me many, many years ago. But all I have is daughters. So i'm I'm well past the being you know weird about being in Ulta. Like I have to go, you know, like he's so good. All think in a wig or something. Yeah. think you go Oh, I would love to see James in a wig. That would be great.
01:08:23
Speaker
Uh, what's a country you'd be okay never visiting in your life. Oh man.
01:08:32
Speaker
You know, I forgot what this country's called. It's like, oh, you know what? It's actually part of Canada, but it's not really Canada. It's it called Baffin Island. I actually went there once on ah on a test flight to test the coldest, ah the the airplanes in cold weather. It was cold. Like I'm from Chicago and it was freaking cold. ah And so I'd be okay never, not a country. It's like a, I don't know, province or whatever. ah Baffin Island. And I'd be okay never going back to.
01:09:03
Speaker
I think we need to hear the story. Sorry. Sorry. If anyone from Baffin Island is listening to this test flight for. Yeah. In flight Wi-Fi. We're testing the antennas to see like how our go by systems are working at Google. Yeah. gay Yeah. Yeah. OK. No, no. Well, we're drilling teeth. We're seeing how yeah in the plane. You don't even need a pilot.
01:09:25
Speaker
you channel setting for this one yeah it's so cool We don't even need to knock you out for this. up All right. This might, I don't know. Maybe this one will work or not. If Tupac appeared before you right now, what's the first thing you would ask him? Oh man. This, uh, this is a good one. Uh, what did you think about the current state of the rap between Drake and Kendrick Lamar? And is ah Kendrick Lamar really his reincarnation? It's the, as he claims to be i think it george peting like it's kind of like a thing. So that'd be kind of cool to see.
01:09:56
Speaker
I like it. But now you gotta ask, is this the hologram Tupac or the real? too Oh, yeah that's right. Hologram. because They did the hologram. Your turn. Well, yeah, I'm still, I'm still thinking about the hologram. ku co I'm just like, Oh, black beans or wheat fried beans in your burrito. Oh man.
01:10:22
Speaker
I love me some refried beans outside the burrito, but I think in the burrito, I like black beans. It's weird. It's kind of like ah a handy dooley or whatever. Like and they love tomatoes, but they hate ketchup. How do you think that is? It's like, I love black beans in the tomato, in the burrito, but outside refried beans and fries. Oh, it's so good. It's like, it's like the gravy on a turkey here for from Mexican food. All right. Um,
01:10:52
Speaker
Do you Instagram your food? A hundred percent. The camera eats first. um Okay. When eating a formal meal with multiple forks, do you start from the outside or the inside?
01:11:08
Speaker
uh i plead the fact no i i outside outside in i think is the way it goes uh i forget though yes that is correct i think is it outside in i thought it was inside out nope it's outside in so fork dinner fork and then dessert fork we're looking this up all right i've been here my whole life it depends on how many forks you have but yes it's outside in we don't we don't do that stuff in good tucky If you eat forks, that's why Kentucky fried chickens figure looking good because we don't use forks. That's where they got their slogan. right Yeah.
01:11:54
Speaker
Although who eats chicken with a fork and a knife anyway, like, come on, fried chicken. Oh my gosh. I grew up. So obviously I had a very etiquette based upbringing and it was, um, yeah, watching, watching somebody try and eat chicken off the bone with a knife and fork is painful.
01:12:20
Speaker
It's just, I'm like, really, just really, it's too animalistic to pick it up. Like, Oh my gosh, way too, you take this way too seriously. I think that was the final question of the lightning round. Did I make it? You made it. You still have a pulse. This is still real life. Yes. You've, you've been struck by lightning and survived.
01:12:49
Speaker
ah So Joe, do you have any final closing thoughts or predictions for the new year on on platform engineering? Yeah, I think generative, i know I'm not kidding. I think generative definitely be something that it's going to get some focus, obviously, as we were in the industry, can't even dodge it on our personal lives, let alone in our professional lives. But I think really, it's, um I think we're going to start seeing um a build a oh taking of the platform engineering to those ultimate ultimate platforms, alternative platforms, it's not just going to be building it for one
01:13:25
Speaker
thing anymore. It's not just going to be for Kubernetes. It's evolving platform engineering as a whole for a different business lines and things like that. So I think that's going to be kind of the the focus of 2020-25. I see a lot of obviously AI, but I think a lot of focus in hybrid multi-cloud environments. Yeah. And you know, looking back at our Our previous episode, we did predict that, you know, platform engineering was going to be big in 2025, if I remember correctly.
01:13:57
Speaker
here i is I know it was on my list of, of um, or maybe it was a shipper or skipper that I talked about platform engineering. I think it, it's an absolute must have and will be as we, as we add more, more ways your code can go. Um, I do see a need for.
01:14:18
Speaker
being able to write code once and have it go multiple places. Well, I mean, using the gen AI example, right? I mean, look at what's going on right now. We're, we're basically making AI kind of somewhat of a commodity. We're asking different models, different languages, and really what's turning into is, oh, that one's better for reasoning and that one's better for mathematics and that one's better for this. And now we're starting to see like these worlds of multiple models being used for different things. So it's almost like, just like we're talking about building commodity and like, Hey,
01:14:48
Speaker
whatever we get, we're going to be able to figure out what, you know, kind of figure that path out.
01:14:55
Speaker
What about, um, I know you're involved in a lot in the local tech scene there in Chicago, any, any, or anything
Chicagoland Tech Community Shoutout
01:15:01
Speaker
else you wanted to do. Is there anything you wanted to plug or share with folks? Or i give a shout out to anybody.
01:15:06
Speaker
Oh, yeah, sure. I'll give a shout out to my local Chicagoland community. So all the Chicago meetups are great. You know, we every year we do a Windy City Dev Fest. So be on the lookout for that for 2025. It's kind of a, I don't know how to describe it, ah you know, pandemic kind of changed a lot of in person meetups. So ah What we did in 2023 was we just basically reached out to like 12, 15 different meetups and we said, hey, we got a venue and we have 12 rooms. You want to just run a giant meetup? And it was like the most disorganized, awesome hack fest I've ever seen. ah We had people with Python, data engineers, Ruby engineers. It was like people were going to talks that they probably never in their life would have attended. And out of it, the positive experience was like, this is great. I just got to meet a lot of people that are actually working in the industry in Chicago.
01:15:52
Speaker
Remember that, hey, you know, despite us working in different industries, we're all, we're all ah you know, the world's small and and so is the city. So um support your local Chicagoland communities, support your local city communities, not just Chicago, um and and focus on that. I mean, I think that's what, at least to me, gets me motivated, excited, um and make your city great. That's just to be a good person overall. Nice. When is the DevFest this year?
01:16:19
Speaker
Uh, uh, the last one nearly, uh, got the best of you of how short of a plan we had. We planned it in less than a month and we turned it around pretty well at the Google office. But, um, we usually do it like September, October ish around there. Um, but but we'll, we'll probably announce it sometime in, uh, in July, June, maybe a little bit earlier this year. Very cool. Yeah. I, I, I spoke at a,
01:16:43
Speaker
the Chicago Java user group. And it's a good town, man. I hadn't been in Chicago in a while and I got to go out and see the city. My dog is barking up here. But yeah, it was a good town. I really enjoyed it. Chicago is a great town. Yeah. The Java user group, the Kotlin user group. I mean, Mary and crew are doing a great thing over here in Chicago. So definitely props to them. They're doing some awesome stuff. I love it. I love working with them too.
Aspen Group's Engineering Opportunities
01:17:08
Speaker
um And then ultimately, I'd just like to give a shout out to um the Aspen Group, ah if you're ever looking for a career opportunity um you know based out of Chicago, based out of Syracuse. um you know We're always looking for great engineers. you know Specifically, I'm looking for SRE and to deliver engineers. So if anything I talked about is exciting, check us out and we'd love to have you join the support.
Conclusion and Call to Action
01:17:26
Speaker
All right. So with that, I will wrap this episode up. That concludes this episode of the Forward Slash, where we lean into the future ah of IT. t I'd like to thank our guest, Joel Fisallo, for talking with us about platform engineering. I'd also like to thank my beautiful co-host, James, and all of our production staff that make this podcast possible.
01:17:45
Speaker
Make sure you subscribe to this podcast for future episodes. And if you'd like to reach us at the forward slash or have questions or needs for platform engineering, you can email us at the forward slash at coliberty.com. I'm Aaron Chesney. Thank you and stay curious.