Benjamin Johnson is the CEO and Founder of Particle41. He joins the show alongside Jonathan and Will to talk about his company. He starts off as he shares his journey in establishing and maintaining Particle41. Additionally, he talks about how they use DevOps as a service and how they meet their client's needs. 

Hello, everybody, and welcome to an amazing episode of the adventure and develops Podcast. I am your host for today, Jonathan Hall, and here in the virtual studio, I'm excited to have Will Button,

 

Hello, everyone,

 

And we have a special guest today, Benjamin Johnson, Benjamin.

 

Hi, guy, it's great to be here.

 

Glad you made it. So you're calling in from the Dallas area. If I'm correct,

 

That's right.

 

Go cowboys.

 

Here we

 

For

 

go.

 

sure.

 

Yes,

 

So before we dive into the topic today, Uh, Benjamin, why do you give us a little bit of a introduction? Tell us little bit about who you are, what you do and and maybe how develop fits into your life.

 

Yeah, sure. so I'm the co of article forty one where full service software consultency, So of course we do customer application development, data science, or data engineering, rather, and then devops. And so if I give a little bit of history here, I did early in my career, I did a travel company, you know, think or bit, Travelosity, Expedia. The brand was called One travel, and in that company I was Ninety nine when we started it, and I had a rack and stack my own hardware. Like who does that And you know, also, you know. not many of my peers have experienced something like that where we were having to run our own B, g, P. routing to make sure we had multiple sources of internet. Uh, the founder of that business wanted to build their data in in like small town, West Texas, So imagine a store front building with brick inside a brick and two fifty white Servers running, You know like an Alton load balancer, So I'm super dating myself, M. But we learned a ton right. We learned about how the Internet really works at A at a super core level, and then from there I went into a travel media company. Um, and you know we're selling adds between these destination guide websites and H and Top line travel advertisers, And you know we Had to adopt the cloud because the data profile that we were getting like you couldn't buy hardware big enough fast enough to process Im pression level data off of the websites. So was an early a w s. adoptor M, and then repeated some of the same types of technology with a financial brand called Investing Channel, That's still around today, Um, but just had to embrace Devops, because, even though the business was say twenty thirty million and total Web, we'd have Spend a pretty decent percentage of that in online internet technology to deliver. You know the best, adds to the best people, and process all that data and make real time decisions on on those ad. Coles. So you know I was using Hedoope the first day it was launched. Definitely

 

Uh,

 

trying to run

 

uh,

 

elastic data notes and all of that and just feel all the scar tissue from the early cloud adoption, and then back in two thousand Of my partner, a friend of mine, who you know I've been working with throughout my career. he and I went in to business together, and we started article forty one to really take the customer application development that we were encountering the data engineering stuff we were encountering in the devoposan, kind of pull it together in a complete kind of engineering in a box consulting company.

 

So the big question I think everybody listening is wondering about is how did you choose that name?

 

Okay, so yeah, so I love that I love this question. So part of L. forty one, If you go to the forty first Element in the Periodic table, it is Neobium and neobim. Not something I wanted to call the company, but I don't know. Growing up in the eighties there was this junk jewelry that was like a pop medal jewelry and it had like an anadyzechene. Well, that was niobium jewelry. That kind of rainbow anadized medal, Um, and so super popular in the eighties, but in the medical, in the medalergical process, making alloys and molding medal, they will combine amounts of neobium with steel with the steel alloy to make it stronger more flexible and look a little better, because then you could sheen it, you could anadize it and give it

 

Uh

 

color.

 

huh,

 

And so when I learned the use of niobium particle Forty one, just kind of clicked for us. And

 

There

 

so

 

you go.

 

We ran with that name.

 

Cool.

 

That's clever. I like that.

 

That might be my favorite naming story I've heard in at least in a long time, maybe ever

 

You guys are too

 

Seriously

 

kind. Yeah, but we think when you know, combining or re essential elements and core values with a company, we can help you be stronger, more flexible and maybe even look a little better. Make your products look a little better.

 

Right,

 

Nice.

 

shinier.

 

Yeah,

 

So is it true what they say that everything is bigger than Texas? Is that apply to devhops To Is devhops bigger Han Texas?

 

So I've been super involved in the. in the past. I've been super involved in the community. Her devopstats d f. w is definitely a stellar event Keeps getting bigger and bigger every year. Um, and yeah, everything is bigger in Texas. Um, also was a co facilitator of the Devops live. I think we're starting to meet again. Um. and so yeah, the Dep community as been great here. You've had a lot of recovering financial Stitutions. when I say recovering,

 

Hm,

 

like recovering from the dinasortof t Ges, and getting into debopsoeyoue. had capital one is here. They've been on the cutting edge of moving into cloud for a financial company, and then we have a lot of telacom stuff going around, so there's a good community here that leans, leans to enterprise.

 

Yeah, Dev, Austin is a big.

 

That's a fun time for care.

 

It is. Yeah, I've been there a couple of times. I actually grew up just about halfway between Austin and Dallas, and so I usually apply to Davos Austin every year to speak, and just turn it into a trip to see my parents. I haven't been to Devpsdfvw yet, but I've heard that it's on par with Devops Austin,

 

Yeah, we have a good time.

 

Right on.

 

Yeah, we sponsored this year. we had a table and you know, talking about, talk about our services. We had a really good time. They treated us well,

 

So what kind of clients does Particle Forty one mostly work with? An? You just talked about some of these sort of recovering financial companies. Is that who you work with or do you work with a different sort of demographic?

 

Um, I would say we're mineprise. You know, we've worked with a couple of folks. we've worked with ration. In the past, we've worked with some large enterprise payment process or type companies, but we've of course done. you know, startups wanting to go to market with the next greatest mobile app? Um, you know, so you know, there's that game between that enterprise level. We'd love to do enterprise clients with our Vos teams. But what we're finding in that miniprise is you know what kind of devups are they ready for? And how can we meet that need through kind of a more product service? So we're really have focused on making sure that the set up is super smooth. So do we have all of our terriform libraries? Can we quick launch them? Um, you know, get through that set up. get them to pipe lines and then figure out you know what are the needs from there And most of our engagements were come Inning, the development, the Devopson Project management, Q. A. You know all the services to have a complete team.

 

One thing I hear lately, especially on Linked in. I think he's been a paid advertisement on this topic lately, The mem that devopsis, dead, long lived platform engineering. What's your take on that?

 

Yeah, so I love this discussion, I think, because we see varying maturities around the enterprise, so that you know if you're getting to be a larger organization, you're motivated based on your scale, to you know, do things properly, So you go through you know, several iterations, and I think if I can go back a little bit into the story here where you, you're racking and stacking your own hardware. when you Come from that you need these teammates that are just all about them for structure. But now we've gone to the shift where you know a lot of that is automated and we're just seeing a lot of trouble, windtaninitial, higher up discussion. So the discussion with me to establish the team is really about preparing the pipe lines, and we are successful in delivering that, hopefully quicker and quicker, as much as we would seem to be repeating ourselves. And and then we get to this P. Where they can't quite let go, and of course, financially, that's great, right. You can't let go of some folks that are on my team. I love that, but

 

Yeah,

 

we often see it wrapped in some less than create activities like blame, like Okay. This pipeline didn't run Okay. Well, was it code drift? Okay, When the code was edited, where the deployment tested again, and where the developers able to participate in that, since they were the change owner, And we see all these kind of conflicts where they can't quite let go. of What what I think the mem is really talking about is this new form of cloud ops that devots sometimes get pigeon holed into. That's the terror form code. We need the Devops guys to touch that, and you know we're just constantly trying to counsel through that increased understanding that the developers need to have with some additional parts of the ecosystem. One, ten,

 

Yeah, sure,

 

So about that,

 

Yeah, that's all right. No problem. this is the part where our heroic editor makes this sound great.

 

Yeah, thank you very much.

 

Yeah, that's

 

So

 

really

 

when

 

cool. Yeah, go ahead, will.

 

I was going to ask you know, in my experience, Deavops is very like. it's a really broad term, and working from one company to the next it's It's very different, and in many cases the only thing consistent in deavops from one company to the next is the fact that they're both using the same word deavops. How do you producttize that across multiple clients and keep your your team feeling like they have some level of sanity and control.

 

Yeah, that's a good question. You know one of the metrics I'm really proud of in my business As our level of attrition. it's been dangerously close to zero just over this entire Youkncovid post Covid, the Great reset, A great resignation. whatever you want to call it, We've actually seen our clients suffering with a higher amount of turnover on their teams than what we're experiencing, And I think it's because working for Botique serve This firm, you're able to focus on your craft, and even if you're engaged with a particular client you're able to hear about and get exposure to the kind of the Devops philosophy around these other clients. So you're either like Wow, This, this place isn't comparing in a negative way. I need to help it improve. So you're you're offering Some may be more advanced council as a Devops engineer to them to kind of pull them along, Or you're in some place that's you know, Got there. E, k, S clusters all orchestrated and it's awesome and they're right on the edge trying to think of their service messed options. And and you're sharing that with the internal hive mind a particle, and you're able to get you know, Really experts as currency internally And we've really you exercise that point. And so I think that's why we've been able to really, you know, hang on to people, and then the idea that we're exploring more ways to sell Kind of devops as a service, producttize and patternize the practice of pipeline development or platform engineering. Um, and then recognize those different levels of maturity and pull people along. I think that's made for a really fun time for all of us.

 

Yeah, it seems like that's a great mix of getting the entertainment value out of your job Because

 

Hm.

 

you get to you don't get stuck doing the same task day after day and you get to see all the different aspects of it

 

Yeah, for sure, and then I think the antithesis of that is in a corporation you're probably more aligned with, like the stock price or the revenue, and as a devops engineer, you care about those things because you want to understand the vibrance of your company. Your ou know professional. So you definitely have a business mind and you want to know about those things, But it's not front of your mind what you care about. You care about, You know, advancing your career, solving those scale ability problems and knowing that you can advance that devots posture you. Those are kind of high of you know, high on your mind, And so I think when when you're in sight of service come you really get that difference in purity of focusing on the craft, productizing and paternizing repetitive things that we're seeing across clients. and you get to invest in that rather than being, You have to go to that next all Hands meeting about how we're going to raise the stock price where you feel so indirectly correlated to that in your role,

 

For sure.

 

I'm curious how how you do Devopsis service, because you know I can. I can imagine. in fact, I've seen this done poorly at some companies where it's It's kind of just basically devopsis their silo that we off shore, And you know once devopsinsi, is it sorry? once devopsis, in a silo, it's not devops anymore. You know, Since

 

Right.

 

devopsis, all about tearing down silos, So M. how do you sort of address that that ten And with companies? And do you have companies trying to put it in the silo? And how do you deal with that

 

Yeah, so I think we have to start talking about the role of and this is why I think the platform engineering conversation is so healthy, Because we're able to use that vocabulary that that. there's some good writings, Really good medium article that you can kind of use to to kind of educate yourself on the difference, and you have to express that like we're responsible for the pipe lines. But how can we empower your developers? And so we definitely have two very distinct pricing models. Like you can have a dedicated individual or dedicated individuals on your team to help you accelerate that pipe line. But then we're desperately seeking to try to evolve that back into the team. Get the developers to self serve. Make sure we've provided good documentation, Um and M. but at the same time we all Want to be there for the next time. additional pipeline is needed. So if they are sprinting towards a particular productal and with that productal comes a new data architecture or whatever, then they can lean on our team and we go into more of a fractional pricing model. So rather than paying for a full time dedicated person, we have a lesser than pricing. And then, also, since we have people in the U. S, we have people in Argentine, and we have people in India. we're also willing to help them with some of the S. r Re. roles like the full time monitoring. And so we're really. I guess I use the word exploiting both the the It part of the misconception, which is not really Dev Ops, but the Ops portion, and then also really trying to define why you would want to accelerate your biplane development so that you don't have to you, so we can come in and do that, kind of give you the keys to the kingdom. help you self serve, But also, if you need Going support, and you want that insurance that you have somebody to go to in times of need and get some of that S R type support, we can do that, too.

 

Nice? really good?

 

So I just read an article earlier this morning. The kind of a line with what we're talking about here. It's from David in a Mir Hanson, who's the founder of Ruby on Rails and

 

Yeah,

 

C to thirty seven signals. Yeah, all of that stuff. his article says We stand to save seven million dollars over five years from our cloud exit. And so they're going to a data center racking and stacking their own servers. And when I read that article, I mean, I just broke out until cold sweat, because I remember, I remember driving to the day to center in the middle of the night. You know, going in replacing hard drives and pulling servers out of rags. And like you mentioned at the beginning of of this episode, you know setting B, g P routes and all of that stuff and I don't miss doing any of that.

 

Right,

 

But and I'm not really sure that they're going to save as much as they think they are,

 

Yeah,

 

But What? what are your thoughts on that?

 

So any time I would do this calculation and mind, I think David d h. H has a different problem that he's trying to solve than the average business. So if he probably has a very static equation of what, What kind of hardware? What kind of compute he needs to run his business right? I mean, he has an established as company, so his revenue is at a certain place. The amount of com He needs. He knows what that is. I. I think that's a lot different than an up and coming customer that's looking to scale. So if you're a start up or your minoprise, maybe you're even an enterprise wanting to do a new thing. You know at the beginning that that thing is going to be low. Your usage of that new thing needs to go through the adoption curve. And so the cloud is super great. You can't do this equation. Like what if I had to go ask My board for you know, five big beefy servers right now, I'd be asking for probably hundreds of thousands of dollars to have an infrastructure that I thought would last me the next five. you know, five years.

 

Right?

 

And so I'd go ask for this big chunk of money and they'd be okay. Well, what are you going to get? Well? this is new. I'm trying this thing out. I mean, I'm not totally sure that conversation goes very poorly right Because you're asking.

 

Yeah,

 

you're asking some of you to take a risk with little to Guarantee of return. Um, and that's where the cloud just totally accelerates, so I think his context is just very different here. Um. it makes for a good provocative article, but he is trying to solve a very known problem. and Um, you know the cloud is there for these unknown problems. And then You know, I mean, God forbid he have a decrease in volume or a big shift upwards. You know, he's going to have to absorb that one where or the other right. And this is where the cloud is. Just you know, really excels. and I don't ever try to convince a client that the cloud would be cheaper. Um, but in most of my

 

Hm.

 

calculations I've seen like a twenty four month, uh, you know, break even, Um. on that initial Vestment to win, I think the cloud is starting to edge up, but then the good news is if the cloud is edging up in twenty four months, that means that you've been successful. you've had to scale up. you've had to turn on the. You've had to throw some hardware at the problem. And so that's probably a good problem to have. Um, so yeah, that's my take on that as I think he's His context is different than the you know, the eighty per cent case that we see in the market.

 

I

 

Right

 

agree.

 

on.

 

My biggest complaint about the articles that he's written on this topic is that they, they don't make that clear. He kind of almost preaches as A, and a lot of people interpret his words as gospel like. Oh, if

 

Yeah,

 

D does it, we all have to do it And that's

 

Sure,

 

like you're not Google and you're not d. H. H. Take that into account

 

Yeah,

 

when you're deciding how to run your own business.

 

Right. Yeah, you know, I did think that he. history Has you know, they wrote the remote book

 

Yeah,

 

about working remote. and man, did they get a winner on on that one. Just with the way Covid played out, You know,

 

Yeah.

 

Yeah,

 

you couldn't have been more lucky to write that remote working book, and then a couple of years later, have that happen. Just have everybody forced into remote situation, so they've definitely, you know. I was a Ruby developer for a long time. Um, so I'm appreciative of the contribution Rubyanrls made, but yeah, you know we got shocked jockies in every industry right, not just not just in radio.

 

Yeah. fair point. So do you focus on clients in our new target clients in the Dallas area or do you go anywhere?

 

No, no, we've been. We've been remote since the inception. Um, we have clients on the West Coast. We have clients on the East coast. Yeah, where we're completely location agnostic That we do generally only Servius space clients.

 

Okay?

 

But yeah, we're about a hundred people. sixty in India, about twenty in near shore, Twenty Us. and mostly us is design talent, project managers. That kind more of the high level consulting in the Us.

 

Got ya right on.

 

Do you have any fun stories you can tell us about a particularly challenging? Maybe, I don't know if you want to. I don't want to ask you to tell us a client's name, unless you just know that you can, But you have an interesting story. Develop story. You can tell us that it's funny or encouraging or anything like that.

 

Well, I don't know. Maybe it would be fun to talk about those racking in stacking days. I feel like we were doing devots before Patrick Deb coined it Davos. We were doing some

 

Hm,

 

flavor of that, but in that store front we had the two hundred and fifty white box servers. Alton loadbalanceers, I don't know if you guys ever had to tinker with one of those.

 

No,

 

I

 

Hm.

 

think

 

al tion.

 

M. Alton got bought by North. Tell, so you might have had to mess with Nort Load balancer, Uh, but we were, so we had to pay local loop to get internet from Dallas to West Texas to Odessa, Midland. We had a brick inside a brick building with a deal generator sitting outside and we were running cold fusion, and then

 

Uh,

 

some

 

uh,

 

some C plus plus extensions behind that that would actually interact with the travel S main frames, so we're getting the pricing data off the g d s main frames so you could price the tickets and show availability and all of that wonderful stuff. But we were serving it through a cold fusion back in and it was not scaling. It was shooting sparks like horribly, and we ended up landing side step, which is a big travel brand and Kiak, And so we just were absolutely shooting sparks Booting servers all the time. It was really really rough, and then the so called us because somebody was having trouble shopping on a fifty six C modum and

 

Yeah,

 

it became the biggest issue that they found out that they still had a good percentage of users on fifty six modum Um. And we found out that it was some kind of buffer issue in the end, But this discussion between the hardware guy, trying to figure out how to pack it, We're flowing through our own upper tier infrastructure. Uh, to figure out why, like half the page, A quarter of the page was loading on fifty six k modum and then just stopping. so we had to optimize the buffer for these fifty six kmotums And just I think, the guy on the infrastructure size name is Chris from. He's He's a Devots guy now and he was my hard R counter part and I was kind of responsible for the software. He was responsible for the hardware, but we still talked to this day about. You know how we approach like C. I, D.

 

So

 

We had scripts that would run and deploy all of the colfusion Kcode to all those different machines, and we were doing some really interesting stuff to allow the developers to self serve. And and so yeah, just it was awesome to be kind of in the sweat shop with Hardware counter parts. And now you know, Hopefully those guys are like some of your best s r s right now that have had to go through those war stories and have the you know have the scar tissue to come out of those battles

 

I love those old stories back when Full Stack had a completely different meaning.

 

Right? Right?

 

Right.

 

For sure?

 

Was there a big um presence Like Odessa Middlin is like primarily an oil town. Was there a big presence for technology when you were there?

 

No, all my teammates on my local teammates came from the local. like I, T. T. Technical institute. That showed some

 

Hm,

 

high aptitude. We eventually converted everything to like a dot net stack in early C. sharp days. Um.

 

Hm.

 

we're able to know the the original founder that wanted to do everything kind of hyper local, you know, right there in Odessa. He eventually moved on. A new guy was hired That I reported to, and we ended up moving to like a Dallas, Colo, and having fifty per cent less costs. for you know, way better service. you know, having twenty four hour hands on support inside the data center, You know is phenomenal to retire all those all those white box servers in this kind of make shift data center, But I just super value the learning because you know, I think Gave a real full full stat context to problem, so you don't tend to meet like my team. Its, don't tend to come out of college with as much generalism as what that kind of school of hard knocks you know, gave myself and my peers, you know, kind of kind of seeing the whole, You know a part of the stack that you just don't see anymore. Um, occasionally, you guys probably run into piers where you're like. Well, we need to organize the siders and they're like. What's a Cider again? Network Apology? You know, you kind of have to clear that up, or as the last time you had to work with a V. And you know, so you appreciate being able to understand some of these topics when I'm a. You know, I'm a code by my background.

 

Yeah, I honestly think that that's a big part of our role as we're getting older. you know, because I'm ready to admit it. I'm old now, but I think there's a lot of like contextual knowledge, like things that we did because there wasn't an alternative back then that that's a solved problem, But some newer people who are just getting started don't realize that that's a solved problem, and every once in a while I come Across someone who's trying to resolve that problem because they didn't know that it's a problem that's been solved for thirty or forty years, And so I think that's a large part of our role is senior engineers or senior managers in this industry is to make sure that we're sharing the knowledge. I almost like in it to you know, sitting around the camp fire sharing stories about the tribe. You know, so that the future generations have that knowledge to build on Op of, rather than re inventing the same stuff.

 

Yeah, for sure, I think my first chef project was, you know, had so many anti patterns in there. I was pulling. You know, I was pulling in the source directly into my chef. Ripe, and then you know, Sausomthe the Birk Shelf way, and I was like Damn it I've done. You know, everything I've done to this point was wrong and you know, had to go

 

Yeah,

 

spend

 

M.

 

a week end refactoring

 

hm,

 

my chef code to be more best practice. Um, But I think yeah, we have those hard hard knocks so that you know if we saw a friend picking up something would say Hey, you know, this is how you organize it. This is how you keep from repeating those same mistakes. Make sure you look at this article about anti patterns. Um, and it's cool that you say that will, because I think we'll see a lot of that with the coding by chat. G, P. T. like. Well, I got this. I got this off. It's going to be even beyond just going in monkey patching With stack overflow to be like, Well, hey, I got completely running code, but I don't really know how it works. I just know what to ask, M. C, g, P, T to write for me, which I, you know, we encourage people to be productive and accelerate themselves, But you still have to know how it works. You still have to know what you're putting down and committing to the code base,

 

Yeah, absolutely. the fact that you didn't write it doesn't absolve you over responsibility of what it does.

 

Right right? And the guy that sees your code next is a serial killer that knows where you live,

 

M,

 

Right.

 

M.

 

Sir,

 

People forget that and I think you have a lot of the same responsibility, even just pulling in a dependency. You know you

 

Yeah,

 

That's

 

can't.

 

right.

 

You can't trust that that dependency is safe necessarily, so

 

That is probably on aveopsecurity. This is reached a really high level importance in sight of some materials we have, so in some of our play books, cloud solution, work books and stuff, We're really talking about that static code analysis as the difference between being like a devopomaturity, level two and level three because you at least care about the the supply chain pipe line that you're pulling in, and Um, love all the work that has been gone into Or cube, and some of the other static code and check marks. Um. Some of those tools were really encouraging clients to incorporate those sooner than later. Um, because we're seeing. Uh, you know when you have the Neo for J thing come out like people who have those things already implemented. That was great because then they could go Okay. Well, we need to test all the you know. We need to deploy all of these pieces of software, update the minor version and deploy them out. I can't tell How many clients though, Um, had like a single Jenkins box and everybody went to go fix the Neo for dependency all at the same time, and they realize we've never done this. we've never like. like, out of all the code we have in our enterprise, most of it as static. So having to go through

 

Hm,

 

something like that, you know, it's first awesome that you had some kind of sad a code analysis to know that it needed to be addressed and you had that inventory some people didn't even have That They were totally rummaging through the co juncture to figure out what needed to get edited. But then the people that did know that went to go fix it quickly and realize that their pipe line was not in any way shape er form.

 

Hm,

 

So given your experience with large number of different organizations and different types of applications, what are some of your go to favorite tools to use in the death up space? Ones that like, whenever you show up, you're like. This is what we got to roll out first, and this is my my favorite version of it.

 

Yeah, so I mean, we're definitely heavily invested in the Hash Corp tool chain. Um. We're definitely going to recommend manage services over any kind of roll your own. Um, And We have not necessarily specialized on any particular framework or coding standard, but we love frameworks, so we're definitely, If somebody says hey, we're kind of a Java shop. then we might introduce Spring Boot, and Cotland moved them away from the overhead of Java, but still give them that spring framework experience that they want. Um, but we're definitely going to say hey, let's do a framework because that's easy for me to manage best practices against. And then You know, see if they have any kind of other language leanings or whatever going on in the organization, and we'll kind of roll that way. Definitely agile. you know, Scrum methodologies. In terms of execution practice,

 

M.

 

we bring a lot of terra form with us, so you know we have terriformplaybooks and terriform moguls that we know and are trusted, and are fairly opinionated terms of intplementing, logging, and all the permission sets, and what not, Um. we are starting to Recommend Uh, vault and Um and boundary for kind of a zero trust posturing, But these are. these are still being kind of like explained to clients a lot, Especially boundary seems like a kind of a form of over kill to a lot of C. Tos and C is Um. But yeah, we're trying to really keep it simple. so, but I would say we're kind of staying away from cloud formation as much as we can, Trying to no role, decent opinionated play books with terror form an eble, as kind of our configuration management tool of choice. For some of those things you just need to orchestrate with ancerble, and then we're

 

Hm.

 

pushing cubernetties as much as possible. Like most everybody o K, S. And you know, climbing the tool chain with that we do see some of the startups

 

So

 

that you know. Maybe having some C. S is more Table, and so we can quick launch an S cluster to get a nap up where we can. Our set up time for K. S. is probably you know, two weeks,

 

Yeah, Whenever it comes to implementing U, K, S or any cuberneties. How much how much control do you give the developers in their ability to self serve on that like, do you have them locked down to specific name spaces? or do you just say here's the cubernety thing. Deploy what you need to.

 

So we like back stage. Um,

 

Oh, right

 

we

 

on

 

like the like the Argo Cd, So we show them how to use that. We said them, But we own the Cubernetti's cluster, So Um, we want to own the infrastructure in a way, and then partner with our infrastructure friends, our managed service. You know, our Ms. P type friends for the cluster itself, But they own the pod. They own everything surrounding thatpocation and we show them where to access Bridge configurations. Um,

 

Right,

 

yeah, I think the setting up of a new application needs to get shifted to the developers as soon as possible. That's where we've seen the most like Rub. in terms of okay. Well, it's a new application. Can you guys to it? and then the dialogue is is Okay. Well, we've given you new application documentation. Here's your Argo Cd. You know stuff. Here's how you set up the pod. Here's where you need to add these things. We try to kind of do a lot of shoulder to Shoulder work in that early stage, and then one by one the developers will kind of get on the the idea of Oh, I can self serve. I can do this. I set up an app last time and finding those champions within the developer organization, Um is essential. You know folks that have done that learning so you can go. Hey, you know, Jimmy. He's done this before and then that starts to kind of cross pollinate the new app set up.

 

I'm curious to know how you get most of your clients. Are you doing outbound Le generation? or do you get mostly in bound at this point?

 

It's been all referral, so a

 

Okay,

 

lot of my business has started, just you know, doing one good client engagement at a time, and then that client engagement rolls into others. We. We've grown grown very fast by referral, and then most recently nowpartnering with M, s. P. companies that want to stay more centered around the around the Compliance aspects and the It aspects of the cloud, and give separation of duties to the customer out of the box, but they don't really want to touch the applications, And so at modernization, the Devas migrations, We've done a bunch of those. We did. a particular client that had two hundred and eighty eight applications and they wanted to lift and shift that mostly so we were able to deploy cubernetdies and get some of their more modern you know in flow stuff, moved into con Tainers and over into cubernettis, and kind of give them that as their new way of doing things, And they wanted that like this is the future. But then they had so much existing application footprint that we ended up setting up c, c. D for two hunder, an eighty applications. So it's kind

 

That

 

of

 

just

 

a factory

 

took a weekend

 

operation.

 

right.

 

Yeah, you know, just jam it out. I love those problems, though that can be turned into like an agile problem where it's like Hey, with as many people were able to accomplish, you know, twelve a week and you got this many.

 

Hm.

 

So do the math. That's how long it's going to take. How can we accelerate? You know what steps can we cut out? Um, but we did lift and shift some really sorry excuses for applications to like you know, Maybe an application that was a particular way of getting at some data that could have been easily combined with another application. or Um, you know, we questioned the usage of some of those applications.

 

So then a follow up question? Are you mostly selling to C, T, S, or you mentioned C. I, S. Who are you mostly sort of interfacing with at the sales level?

 

It depends on size a company, so we work with

 

Hm,

 

a lot of Cos. We can do product strategy M, so we can do the product development into end, including the road map, the product strategy, The you know, the whole plan of your product launch, and so in those engagements we're working with the co co founder Um. At the Mine prize level, we tend to be working with E, C to Um. or because we do want to have the well round Team. I mean, we feel like we're doing our best work when we have when we're tied to an initiative, and we have developers to have Ops Q. a project manage, are all kind of working towards that same goal together, and there's a little etonomy there with some of the larger organizations we're reporting into like a v p of a particular business unit, So it may just be you know, devopsaugmentation or data engineer augmentation, but I would say those are. Those are more rare that we're kind of just Meting a specific rule.

 

Is there anything else you would like us to talk about or questions that we should be asking that we haven't thought of or haven't done yet,

 

Um, I guess I'm curious about your guyss thoughts about Kind of this idea of the Defopsas a service and the value it might bring to have like an external group, Ople to handle the three oclolk in the morning call. or you know, have you seen that work for any organizations where they kind of out source the Sr function? Because I think that's what I'm really talking about when I talk about. Devops S a service.

 

Hm,

 

you know, maybe a little

 

Hm,

 

pipe line up dates or pipeline maintenance, but really I think it's more the value for in my mind is to have somebody Knows enough about your applications to jump in and s r function. Have you seen anything like that work where s R S kind of done externally, or you know, Follow the Sun application support.

 

I haven't seen it first hand, But in principle I agree with that approach. I mean, kind of all we taking about earlier. You can't just kind of put develop in a box, and a lot of people do that people don't realize that. That's kind of the antithesis of what Devops was originally meant to mean. Um, So you, without trying to get too philosophical and in definition oriented, Uh, I think it's important that as you mentioned earlier, the developers are empowered. self serve as much as possible. There's always going to be a barrier. At some point, You know, if the physical servers on fire, the developers probably can't fix that if it's in a different state. But as far as their application goes, you know the developers should be able to do their own deployment. Their own trouble shooting, probably means they have access to logs or alert or whatever metrics. Observability stuff, Um, roll back and all all that stuff. Developers have access to do those things. then, Yeah, I mean that like I think you said that, you know. that's kind of the idea of a platform team and I think it. I think that's a reasonable place to put that that sort of boundary. You know, if you want to out source that platform team, call it or or whatever, I think that's I think that's fair,

 

M.

 

and I think

 

M.

 

it's completely fair for a company to want to be Say, product focused and and not focused on managing cubernttis. That's fine. That's appropriate a certain scale. It's necessary. you know if

 

Hm,

 

you're if you're a three persons start up, you don't want one, and to have people dedicate in their their lives to managing cuernettis. You want to out source that to some company that can do it for you, at least until

 

Hm,

 

you grow to a certain point. You know, once you're H. H. then then you know you start

 

M.

 

having differnt conversation. But

 

Yeah,

 

Yeah,

 

I think that's a

 

So

 

boundary

 

yeah, I think

 

assumption.

 

it makes perfect sense, although

 

You don't want one and

 

Go

 

a

 

ahead.

 

half people dedicated to Cubernettis, but you want a number greater than zero people dedicated to Cuperneties.

 

Yeah, the point is that out of a small scale, it just doesn't make economic sense to have any person on your small team knowledgeable enough to take advantage of cubernettis. But you might

 

Yeah,

 

want to use Cupernetis if it's appropriate for you, but that means you have to outsource it to somebody else.

 

Right.

 

Yeah,

 

that's That's great thinking. the. Yeah, I kind of lost that thought. No, that's great. I mean, that's what we think is is needed As we. We've come in on the pipe line and then we know we're leaving them in a bad place to just end the engagement. So we're trying to kind of fractionalize that that, after support, Um, just support their evolution. Know forward and still be a somebody. I still have like internal communication. Here's what I was going to ask to one of the things we struggle with a bit. Since it is a pool People right, it's a. it's a team of people. So one thing we struggle with a bit is we have some access patterns that we do so of course you can federate your w. s. single sinon stuff. You can kind of link organizations and say okay, you manage your list of users will manage Rs. But then we tend to want to have vault inside so that we can just provision access on an as needed basis and you can log it, But have you seen any kind of access patterns Because it's a different problem. You know you don't want to be on boarding and offboarding all the time, right? I don't want to pass my on boarding off boarding problem on to the client. So we're We're trying to be creative by having you know, obviously federation through the S account, but then also vault, protecting the actual services. And then you know we're even thinking about mandating boundary. All So so that you know we can provision access on this kind of hive mind. Approach any thoughts on that?

 

Um, A few. I want to go back to the the previous question there, Talking about deavops as a service. I think there is actually a really strong model that you can look at for inspiration on that because the developers love services like like Web Flow, which is a new code building and deployment solution. Hero is really strong.

 

Yep,

 

Verceil is killing it with Next J S. and they're hosting service. And so when you look at why all of those are successful, it's because of what you were just describing that the developers can write their code And then there was a structure defined way for them to get their code into production, and I think those services work great for smaller development teams. And then when you hit scale, you know, I think it makes sense to use something like a w. s. But if you were to marry those two together, so Here's your as account with your manage services managed by us that give you a developers experience like like Verceil, I think you would have to hire security to beat raving customers away from

 

Right.

 

your

 

Sure,

 

office.

 

yeah, the platform as a service S definitely a cool trend. I mean, remember using Engine Yard back in the day for a start up.

 

Oh, yeah,

 

Those are those

 

Hm,

 

I

 

are

 

forgot

 

really

 

about

 

cool.

 

that.

 

Plays the mark up on those was usually the same mark up as what the N. s P is wanting. So yeah, I definitely agree with you that the M. s. Ps will need to give like more of a full service pipe Line to their in customer. I think that's essentially

 

Hm,

 

the bridge that we're building with our play books and our, like, We set up this stuff pretty quickly at this point. Um,

 

Yeah,

 

and then you know, immediately getting into the kind of the the nittygrity of the cloud solution.

 

Yeah, because as you scale you know, it becomes there's a few more one off scenarios that you have to address That you can't just throw out a blanket solution for Um, but going back to office like I, I understand, Really appreciate your The way that you you phrase that you know that you need to be able to on board and off board people on your team without harassing your client every time to make sure they have the right access. Have you had any success with using something like Acta to do S so provisioning across multiple applications?

 

Yeah, so we certainly have, And when we do the dedicated model, that's totally fine, Like they'll get on boarded into the enterprise account. They'll get an enterprise identity in the client's account, but in develops as a service, We want to kind of dictate today. who's going to do that. That worked right. We may want to make

 

Oh,

 

more

 

I see.

 

we may want. so it's more of like You know, you never want to give security access to some group of people like. I'm going to give you an Email. Alis, Go give me you know administrative keys with an email Ile, like wait a minute. What's going on here? So we're trying to think of more of a separation of duties where we can control access. But then we like the vault temporary w, S keys. and we like boundary

 

Hm.

 

for kind of a zero trust counterpoint to the more federated access pattern

 

Yeah, Yeah, that's interesting. I've not. I've not tried to solve that or thought about it, but it makes a lot of sense. You know, if we send Jonathan out to work on a particular task today, it would be nice to know that his credentials expire at the end of the day.

 

Correct. That's exactly the idea that So Vault gives you that, and I'm sure you could do some kind of a w. s thing to you know, a little more native if you didn't want the Ashley Court product involved. But yeah, we like this, this idea of giving situational access for a particular service And then what boundary does is boundary gives you the service. So You know for this task I need to connect to h, r. d, s post cress. Instance, what it will do is it'll mount local host at some port and local host at some port on your local machine. Is the r d s post ress, instance In you know, production or whatever, So it's It's basically an alternative to V, p N. So V, p N wouldn't

 

Hm,

 

be zero trust. V, p N would be like Okay, I give you access. You access the entire You know network topography or in inside the network. So what a lot of organizations are talking about is, let's take v p n away. replace it with boundary, and then someone requests access to a service and then they have to document the situation so that's matching up with like a zero trust standard. And so we're seeing those two as kind of a given a take to letting us control our user group inside of our own. A W. S. Well control those users you know will insure that only you know that we on board off board properly. But by putting them in this so group then they should have to go request these things from vault or boundary.

 

And is that integrated with your ticketing or tracking system to document that?

 

We have discussed it because we think that the opening of a ticket could just do the boundary request for sure.

 

Yeah,

 

and

 

yeah,

 

we have thought about that kind of no Ops upgrade to what we're doing now, But the reason why was asking you is because I'm having this conversation a lot and trying to explain the value of these federated accesses, Um, and then the upgrade to security that the zero trust met Boundary in bolder providing, Um, Yeah, I think it's a. It's a really cool problem, especially in a post Covid world where we're all sitting at home, letting our kids use our, you know our computer to do their

 

Yea

 

Hm,

 

online classes. You know.

 

for sure. Yeah, we've actually been looking. Cloud Flare has a zero trust product that sounds very similar to boundary, and we've actually been looking at that For those exact same reasons. that and

 

Hm,

 

for us, like at some point managing the v, P, N just becomes a huge tassel. You know, Like our environment is spread out over multiple regions and multiple as accounts. Because we're using control powers, we have a bunch of different as accounts. And so managing the v. p N on that is just getting to be a huge pain.

 

Yeah, for sure.

 

Well, we are coming up on an hour. We can go

 

Don't

 

longer, but if there's anything else that we should highlight, we could do that before we move on to pick any other questions that you think we should be asking or other topics you want to talk about.

 

No, I'm cool. I think

 

Okay,

 

it was good conversation.

 

all right,

 

No reason to force it.

 

all right?

 

Let's make it awkward

 

Uh

 

Oh, I can do that. Trust me.

 

Right.

 

h.

 

Usually just having me in the room is enough for that

 

Yeah, All right. Well, let's move on to pick. Um, Will do you have any picks for us today? Do you want to start off

 

I do. I'm actually going to pick Dhahs. Say about saving seven million dollars over five years. We will post a link to it in the show Notes, and I'm curious to hear what our listeners think about that. So after you read it, Um, jump over to our read it form our slash Adventures in Devops, and just completely unfiltered response. Let us hear it all right there. And yeah, that's my pick for today.

 

Awesome? and we should get D H. H on here to talk about it. If you're listening to H. H. We want to have you one as a guest.

 

He might have an opinion on it.

 

Oh,

 

He might. now, All right, I have two picks for the week. The first one is my brand new uhpodcast. It's actually not new any more. I've Ben doing for a month, and by the time this episode comes out close to two months, but

 

So last month

 

yeah, I know so to twenty twenty three.

 

Okay,

 

Um, but it's related to the topic of Devups. It's specifically about Go, and me and a co host are doing a weekly Go news program called Cup o Go, So you can go. Cup o go dot De. That's just the letter O. Not like of it's not cup of go, cup o go dot Dave. you can subscribe in all your favorite podcasting listening tools which everyone you're using right now. you can subscribe in that same one. Um, and the promise is keep up to date with the Go community in fifteen minutes a week, So we do fifteen minutes of Go news and then we usually pad it with another twenty to thirty to forty minutes of an interview or some other discussion. Um, but you can skip that part if you don't want it. If you do Wanted to keep up with the Go news fifteen minutes a week. That's all it takes.

 

Hold up. What's this other co host stuff?

 

Yeah,

 

What do you mean at a co

 

I didn't

 

host?

 

want to tell you

 

Yours?

 

about that, Will,

 

This is how I find out in the middle of a podcast

 

M, M.

 

Jerry Springer style and he's right

 

Right?

 

here. Come on in.

 

Well, it's a good thing to keep a folding chair in my office for justice

 

Oh

 

occasion.

 

Yeah, so I have a co host we met on line. His name is she Mod. I hope I said that right is a Hebrew name. He's in televive, and he's a great guy. So listen to the show. Um, listen to how we get along and once you listen, will you might not be so jealous any more. We don't have quite

 

Whatever,

 

the same report that you and I do

 

whatever, I'm over it.

 

My other pick, which well might need. Now

 

Yeah,

 

I learned about a new brewery here in Amsterdam. I don't know how broadly this beer is shared around the world, but there's a new brewery or at least new to me, and I'm going to butcher the name, but I think it's pronounced brown, Do brown. I'm sorry to any Dutch speakers out there, but that translates to brood by women. So it's a bunch of women to start Brewery, and they have some cool beer flavors and they have a bar. If you're in Amsterdam, you can go check out their bar and try directly if you're not in Amsterdam, like most of you are. Um, maybe you can order it from a local beer store. I don't know if it's if it's exported yet or not, but that's my second pick for the week.

 

Nice,

 

How about you, Ben, and he picks from you

 

Yeah, so I've been reading. Well, We say reading what we really mean is we're listening to an audio book

 

Uh,

 

right,

 

Right,

 

So I've been reading in air quotes, Atomic habit, Um, by James Clear, I think, but it's It's really good, and then I'll follow that up with a bit of a rant, Just a check against ourselves here. I've recently realized that I have This habit of the reels. right? You realize that you're on Youtubereels flipping and five, ten minutes has gone by right. So

 

Hm.

 

I and then I have kids. So I also kind of noticing these. How addictive that can be right. So what was the Charles June quote Like, if we don't make the unconscious conscious, then well we'll be. We'll have failure, but we'll blame it on fate. right, because we won't realize That we're doing these things, so there's a little bit of a rant. I've been talking about it internally inside the company. Just check yourself right because the the habits that cause you to procrastinate responsibilities are have like this anxiety consequence right, We're not doing what we. The advantage of being good time managers is that we can have discipline. but then we get the freedom from that discipline, right we? we get the things that we need to do out of the. And then we can do the things we enjoy too, or spend time with family, And I've just been really like, convicted, or thinking a lot about valueless activity and the difference between rest and recreation, so some types of recreation may not actually be restful. And so that

 

Hm

 

we're tired all the time, we got to kind of look at. Okay. Well, are we really doing rest or we just doing recreation? And so I would put

 

Or

 

social.

 

g. r. p. C. Rest, restful grip, Sorry

 

Yeah,

 

That's right. Um, and uh, yeah, So I think that you know the atomic habits was good to kind of think about that. you know. I've done some the leading of some apse on my phone for periods of time just to see about De programming some of those habits, and just I think the younger folks the professionals we work with they may be having. they might even right, Nize that they're having higher anxiety because of procrastination. Aids. Right and then the weight of the things is kind of stacking on top of me. Have college college age kid and he's talking about his peers. Don't even realize how their time is being spent. They're not auditing their time spent. They're getting behind in classes, and then they got to do kind of a catch up or work off the debt, so to speak, And that's really stressful. to feel like you're in a deficit and have to dig yourself out And so are we doing to ourselves that. so I thought atomic habits was really good. Talks about habit for men, and Um, and then I've been talking to my team internally about this, just so that we can. I want them to have work life balance. It's not about grinding all the time, is just about staying on top of things so that you don't have to feel that pressure and that anxiety of the Um. You know the task that you've put off

 

Absolutely couldn't. Couldn't agree more, and I think that's a great thing to point. out.

 

Book has been in my radar for very long time, but I haven't read it yet, so probably need to edit to my actual reading list now

 

Move it from the backlog to in progress.

 

To this print back lot.

 

Yeah,

 

Right, Um, I will. the restful versus g, R. P. C. M. Joke was really good, so I had a couple that I'll just tack on there. So

 

Okay,

 

you know I tried to date job, but I was just objectified, so you know I had to

 

Uh,

 

had it. Really,

 

Oh

 

you

 

yeah,

 

know,

 

yeah,

 

consider

 

h.

 

what they were doing, But then I tried to date job a script, but I never got a call back.

 

Yes,

 

Oh

 

We need. Well, we need the little. uh, but I'm cheating. Sound of fact,

 

Am thin.

 

you have the control of that, don't you?

 

I got to find it. I'm going to let you down.

 

Uh, huh.

 

Wait, Uh now, Riverside s letting me down. It's got a down load audio file,

 

Oh,

 

Then buffer it. The joke is going

 

The

 

to be

 

moment

 

stale

 

the moments

 

by then.

 

practically gone already.

 

Yeah, sorry. I'll do better next time

 

I hope our editor has that sound effect to put in there. If this makes the final cut there we go.

 

Uh, uh,

 

Sound effects provided by Internet Explorer.

 

Yeah, All

 

Yeah,

 

righ. Well, this has been a great conversation. I've enjoyed it. Thanks for coming on, Ben.

 

Yeah,

 

help people join it, too.

 

I tried to date. I tried to date C plus plus, but he was just way too messy. He had so much garbage laying around,

 

Yeah,

 

I think we might need a new channel for a stand up programming comedy.

 

So No, it's

 

Where

 

great talking to you guys. It was great. Talk to you guys. Next time. I'll time all those jokes a little bit better.

 

Yeah,

 

I'll work on my, on my timing for the prompts here to

 

Uh,

 

the audio prompts.

 

uh.

 

We'll just lay Internet lag. That's

 

Yeah,

 

that's That's always easy to blame anyway.

 

All right, we'll do our new podcast where dad jokes meet computer jokes.

 

Well, I guess we'll just keep talking until we'll hit to that stop button. I can't. I don't have

 

Oh,

 

the button

 

you're

 

so.

 

waiting on me. My bad right to see everyone.