DevOps and Docker Talk: Cloud Native Interviews and Tooling

On this episode, I'm joined
by my friend David Flanagan.

You know him as RawKode on
YouTube and around the internet.

And we've known each other for years
in the Kubernetes and KubeCon scene.

Uh, seeing each other at the
conference about twice a year, at

least once a year for many years now.

So.

We've got together a few times and talked
in streams and just, uh, in videos about

what's going on in Kubernetes, and he's
back to talk about a big launch of his

that I'm very excited for us to cover
and get into the weeds because it's

something that every team deals with
and everyone who's learning Kubernetes.

Has to deal with and we
are gonna get into it.

So what this is about is that the
landscape, the CNCF landscape, that's

the thing that we all point everyone
to when they are curious about all the

projects in the Kubernetes community.

And we go send them off to landscape
that c ncf.org, and then they

look at this giant list of over
a thousand open source project.

Well, they're not all, all
open source, but mostly open

source projects that relate.

To Kubernetes or they provide an
additional functionality to Kubernetes

And other cloud native technologies.

I guess it's not just about
Kubernetes, but the challenge

there is where's the happy pass?

Where are the recommended stacks?

Where's the mean stack?

The PHP plus MySQL stack, like
there's all these, in history,

typical stacks of usable technologies.

I was one of those early guys that was
using node js with Mongo because they

seemed like they both sort of rose to
fame in the early 2010s at the same time.

I was developing a lot of backend
JavaScript apps that ran on

Mongo and learning how to use
NoSQL in times where it was cool.

Although now looking back, I probably
could have just stayed with Postgres.

But the point here is that we
all struggle with this, and

especially as you're learning, you
don't know what to learn first.

And RawKode has created an entire website
that we will talk about through this.

Episode on his preferred stack,
and it's not just a list.

This app is quite in depth.

I'm gonna call it an app because it's more
than just a display of listing projects.

It's got commentary, it's got
spicy takes, and he breaks down

his tech matrix solution of.

Cloud native ecosystem.

The way he sees it, the things you need
to evaluate, the things you need to

skip and the things that he advocates
you check into his advocate list are

the must have technologies, and that's.

What we go through in detail is
each of his must haves and why.

Maybe a little bit of his why on how
that came to the list of must-have.

Kubernetes technologies.

We also cover other topics like
Flux versus Argo CD, the HashiCorp

slash IBM controversy and the
CD k deprecation of Terraform.

content creation challenges that
we've both had because we're both

making courses and YouTube videos
and podcasts and all that stuff.

And he's launched all this on
his website at RawKode Academy.

So check it out and go check out his
resources because he's making fun

stuff and I, I really appreciate the
level of effort he's giving to the

community in his open source work.

Here we go.

David, welcome to the show.

Thanks for being here.

no, it's an absolute pleasure to be here.

Thank you so much for having me.

I'm excited to sit down
and talk tech with you.

Yeah.

we've also talked this year a little
bit about course making and the struggle

of being a full time content creator.

we both suffer together and not alone.

In this content game.

I mean,

yeah, I'm still not sure.

It's a good idea

it's like the whole
DevRail paradox, right?

Everyone thinks it's glamorous.

You go on planes, fly business
class, stay in hotels have meals

cocktails and just travel the world.

But really underneath, it's a
very difficult job where you

have to be face all of the time.

Putting yourself into, the public
eye and social situations that don't

come easy to a lot of developers.

it's not easy for me.

even doing this online being vulnerable
on YouTube your website LinkedIn and

making content, it's so difficult.

Yeah,

I, don't think I'd change it for anything.

It's, a whole lot of fun when you
get to make up the content that you

love and enjoy and hopefully find
that others love and enjoy it too.

the creation process, I think
is the most rewarding thing.

you just shipped a video this morning.

little discussion of HashiCorp and
some stuff they're doing over there, I

haven't seen that video yet, but it's
always nice to click publish, because

now you know it's out there and you could
take it back, but you probably won't.

So it's only forever on the internet.

what could go wrong?

I very specifically and purposely
content creation life, which is now

six or seven years, have never really
had strong opinions on my channel.

I focus on showing people the tools and
how to be successful with these tools,

and I very rarely say here is my opinion
on something the video I published today.

where the title is, fuck you.

HashiCorp dash dash dash an IBM company.

That gives you all the context
that you need right off the bat.

Right.

We're

but I, I mean, I have been
burned with this announcement,

of the deprecation and CDK.

So I felt it was time to make it personal
for once which again, I don't rarely do.

So I'm very

Yeah.

I hope it is well received.

Well, it'll certainly mean more when
it comes from someone who's not posting

that kind of video daily or ever.

Either that or it's the end
of my career, and I'll have a

full time job by January 26.

Who knows?

I give a lot of opinions, but I don't
tend for them to be negative because I

also know a lot of these people, a lot
of times I'm talking about something

where I know someone that works there
and they're all doing hard work.

It's just the company's overall
decisions impact things, and

it's not individual's fault.

So I tend to hate throwing
shade at companies in my videos.

But yeah, sometimes
it's warranted, I think.

I give my opinion on the sunset of
that project and what the next steps

are for people I am angry with the
whole thing, but not at the individual

developers who are working on Terraform
CDK, or even Terraform or anything else.

The same with the license change, right?

I'm not angry the people
doing the real work.

It's the C suites and the executives
on their yachts with their billions

who just get to sit and make decisions
that affect an entire community

without a face in that community.

So why should I worry
about offending them?

It's kind of my perspective now.

It's like, I'm sorry

you don't have the community's best
interest at heart and I don't have yours.

you're, talking in more detail.

I tend to pay attention to those topics,
but there's so many companies in our

field that it's hard for me to really
keep track of who's in the C suite, you

know, what company's situation is like.

I totally forgot they
were being bought by IBM

you're right.

Sometimes you do forget that, IBM
has swallowed up Red Hat as well.

red Hat is an IBM company now,

Yeah.

and around we go.

I think we're, especially in the AI
space, watching a lot of acquisitions

and it's hard to keep track of
who now is part of who, but today.

I reached out to David because we
have seen attempts before at trying

to take the giant CNCF landscape,
hundreds and hundreds of projects.

are we over a thousand yet?

Like of all the things in
that landscape website?

I'm trying to remember,

is.

Yeah.

yeah.

it's daunting for those of us that
live at full time and if you aren't

wanting to buy a vendor's prescribed
stack They tend to have certain tools

that they either bundle or, work with.

So you kind of get a chosen stack there.

A lot of us that have been teaching and
educating others have been frustrated

since the beginning of all of this, the
number of projects continues to climb.

I don't, there's no sign of stopping.

And, we, occasionally see
projects deprecated or archived.

And even if you thought you were
current a year ago or six months

ago, you're no longer current.

And what's possible
with all of these tools?

And we do, I think, tend sometimes
as engineers to just have this

like candy land approach a
little bit, where we want to,

tuning

oh man, this exists, so I
must check it out and go,

support

I might need to put this in my cluster.

Right.

And I think

reach

I'm going to guess you're one of those

my

that

at

years it's been more of
a less is more approach.

year.

very

you put out your opinionated
Kubernetes stack.

tell me about that, like the original
inset, like why did you want to

do this in such a, detailed way?

it was more than just a single post that
says, here's a list of tools I recommend.

I think you and I have something in common
and, that we talk a lot, we talk to a lot

of people, when we go to conferences where
we're in discards, whether we're on social

media, a lot of people
value our opinions on stuff.

as a result of that, you know, I have
a public office I always think we're

people big time and they just want to
talk about what they should be learning.

What's their next step in their career?

What tools should they
be paying attention to?

When I go to conferences, it's the same
again there It's okay, I be using Loki?

Should I be using Meemer?

Should I be using Cortex?

Like what does my
observability stack looks like?

What does my security stack look like?

These are questions people have,
because the CNCF and the cloud

native landscape make a lot of money.

They attract a lot of companies who
build products that solve problems.

there's a lot of competition.

I used to send people to the
CNCF landscape document before

it got to what it is now.

But it doesn't really make
sense if you send someone to it.

There is just boxes and boxes
and boxes of decision fatigue.

not like you go there and you go,
oh, well how, about application

delivery components should I use?

Because you get presented 30 options.

How do you make that decision of
which one is right for me and my team?

I decided there had to be a better way.

I realized.

for the best part of, I think
nearly 20 years, 15 years, been

putting out their tech radar.

And I used to use this all the time back
in 2013 to 2015 when I was, director

of development for a software company.

I wanted to know they were doing the
hard work, researching, kicking the

tires, doing all the, syncing with salt.

Right.

So I trusted them for that opinion, but
I realized that they're not really in

a position to cover something that's
as broad or as specific that I want.

And I know I use broad and specific at
the same time, but Kubernetes, Cloud

native is broad, but it's also a niche,

Yep.

tech radar, like ThoughtWorks?

And then I tried to copy it.

Not the code, but the idea.

And I was like, okay, here's my quadrants.

And I was like, this has to be simpler.

So I sat down with my good friend
Claude, and my good friend ChatGPT

and my other friend, Gemini I
had a conversation for days.

This, discussion was massive, it was
like, what do I care about in this space?

What decisions am I making?

If I were going to build a stack
tomorrow, what would that look like?

What are the trade offs that I make?

And eventually I got to the point where
these discussions really hit home.

are the properties of a project
that are important to me.

Here's why they're important to me.

I made, a database of the things that
are important with the technologies

and then grouped them together.

I didn't want to deviate too much
from the CNCF landscape because

there is a lot of information there.

The only thing that's really
missing is that opinionated piece.

Why should someone adopt FluxDV?

over Argo CD that is not on the landscape?

It will never be on the landscape
because the CNCF is Switzerland, right?

They are there to home these projects.

They're not there to tell
people how to do anything.

And that was kind of the start of it.

I called it the tech matrix because
it is a matrix and that we have,

different criteria and groupings and
everything that we need to balance

two projects against each other,
or two segments against each other.

that was really the start of it.

I was also in a very fortunate
position where the Rockwood Academy,

you know, I've done over 400.

300 videos now, covering a
lot of these technologies.

I already had some of the
database just sat there in

YAML file in a Git repository.

I just had to enrich it
with some more information

and then add my flavor
to each of them as well.

And then sitting down with my good
friends, Claude, gemini and I and

ChatGPT, I was like, here's my idea.

Let's build it into the website
Voila, we, got the tech matrix, and

we can go through it in more detail.

I can talk about how some of the decisions
were made and my opinionated stack, but

the whole point of it is I can update
this matrix whenever I want, my opinion

changes, as new technologies come out
just give people a little bit of guidance.

the opinions that are here should you
be paying attention to our technology.

And I break it down into should you be
just keeping an eye on this technology?

Should you be exploring
it and kicking the tires?

Should you be actively learning
this because it will be a

net positive on your career.

Then into the more of the team dynamic,
is this safe for me to adopt with my team?

And lastly, the advocate, do I
specifically say this is the one

technology that everyone should be using,
and if you're not, you're doing it wrong.

That is kind of where I went with
the matrix, right down to that.

You need to use these tools
if you are serious about

platform and Kubernetes in 2026.

So would you say this is a good
resource, for someone getting started?

is this meant for people that are
sort of, do you come at it from like

a problem fixed perspective of like,
I'm thinking about service mesh.

Let me see what David has
to say about service mesh.

is it more like that

So it's multifaceted, right?

If you are new to cloud, native to DevOps,
to platform engineering, to observability,

any of these things, Any of these niches
within the niche, you can go to the tech

matrix and I will give you an opinionated
handful of technologies that you should

be watching, exploring and learning.

And if you stick to that, I hope you
will get enough value to make the

next 12 months of your career come
by in six months, We're trying to

cut down that mean time to progress.

very difficult for someone coming
into Cloud Native to take a

look at that landscape and know
what technologies to play with.

Should they invest six weeks
learning flux and then go spend

another six weeks learning Argo
and building their own opinions,

I want to speed that up for them.

Yeah.

If you learn that, you learn everything
that you need to learn, and I don't want

to say anything bad about Argo, but it's
just not a tool I'm ever going to use.

So I'm not going to tell
other people to use it either.

it's safe to adopt, but I
would never encourage it.

If you are a team or you've already been
doing Kubernetes for 7 years, you still

need to keep up to date with, the changes.

my tech matrix is still valuable
because as technologies are announced,

as they are released, as they are
evolving, you can just subscribe

to a technology and get updates.

on has it moved from my kick the
tires explorer into my learn phase?

Have I now decided to start
adopting it and advocating it?

you can just get the alerts for the
things that are important to you.

You can just get alerts on get
ups if you want, or you can just

scroll very niche into Flux cd.

the way that you consume it is going to
be different from person to person, but it

depends on what you want to get out of it.

I hope that I've structured it
and composed it in the right way

where you can just get what you
want from it without the noise.

that is the problem of the landscape.

It comes one way, the noisy
way, and it's very difficult to

get the signal out sometimes.

Alright,

have

definitely,

And that's

you're right.

I'm now thinking about my
own use of, the landscape

it.

website and

not something I send people to anymore.

week.

it hasn't been that way for years To
me, it feels like it's only there for

people who are in the CNCF and are who
are aware of a lot of the projects and

are looking for what's new, browsing
through some of the categories and

identifying what we don't know about yet.

But that's a far cry from
implementation or a decision on product.

And, especially when we start talking
about platform engineering stacks,

there's a lot of expectation in what's.

What features exist in that stack
versus someone's just saying,

I, need to deploy Kubernetes.

And that there's a lot, I feel like
there's a lot more minimum decision

amount for a platform engineering
stack, I feel like is, a much higher

bar than I've got a working Kubernetes
cluster that's in production.

so why don't we go through some of it.

I'd love to, get some of your thoughts.

we can read them on the website, but
I'd love to discuss some of them for our

podcast audience and get some of your

if you don't mind.

let's do it.

we'll do an experiment, right?

Let's go to the, landscape, and
assume we want to decide which

service mesh we should adopt.

All right?

So we come here and we're like,
okay, we're looking at projects

and products, side by side.

We need to find where service
mesh is, or networking,

depending on how it's structured.

we see we've got a service mesh
section here, so we can group in.

Now we do see Istio and Linkerd
because they're graduated projects and

everything else is a paid for product.

There are no opinions here to tell you
whether you should use Istio over Linkerd.

right now.

STL for a start says simplify
observability traffic managed security

policy with the STL service mesh.

STL simplifies nothing, that is just.

Wrong, can see how many
GitHub stars it's got.

What does that tell me?

Well, not really much.

then we've got Linkerd, which is
drastically less GitHub stars,

which I think is, an absolute train,
but it is an ultra light, ultra

simple, ultra powerful service.

mesh.

I would agree with that.

I'm a fan of Linkerd.

I'm always going to side with Linkerd.

then we've got all
these paid for products.

when should I build or buy?

When should I own my own service mesh and
when should I pay someone else to maintain

it for me that isn't really covered.

Yeah.

It's valuable for a content creator,
right, where you want to come and see what

random technology will I cover this week?

You know, close your eyes and move
them out and see what you get.

But from

for joining us today.

Please make sure to sign in and
have your registration information.

We've got a lot to be
doing So I'm going to go

We have all of the technologies that
I've ever covered or ever had any opinion

on within the cloud native landscape.

I tried to make it very obvious
what these categories are so you

can expand how to read the matrix.

I explain what watch, explore, and learn
are, those are the most important ones.

right

very easy just to come here
and get what is my stack.

It's everything in this column.

There's no other decisions to be made.

The adopt is a bit broader, and that's
because I made the decision that CNCF

graduated project, it is probably safe
for any organization, no matter how

big or small in the world to adopt.

I am confident that CNCF will
keep that project, active and

maintained for as long as they can.

I have never seen them sunset or
graveyard, any project that has ever

been graduated, I could be wrong, but
I don't think that has ever happened.

They put a lot more effort into trying
to save it if it got into trouble.

I imagine.

Well, yeah.

Plus the conditions to graduate, right?

You can't be a single vendor project.

There has to be multiple people
that are actively invested in

making that a successful project.

So, you know, the likelihood
of one company going bankrupt.

50 50 probably worse.

multiple, within a single thing,
I think it should be safe.

So where do Do you wanna
start in my graveyard?

Do you wanna start on my watch list?

Do you wanna start with what I'm
advocating my opinionated guide to

platform engineering Kubernetes.

What would you like to learn?

Yeah.

So, for the audio audience, I'm going to
mention the numbers at the top, the live

totals where it says skip, watch, explore
at the very top are those your totals

Yeah, so I mean, I

For the audio audience, I wanted to
mention you've actually gone through

and deselected essentially 169.

projects,

Yeah,

that's 169 CNCF projects Correct.

so in total we had 320 projects
that I have stored data on in the

Rockwood Academy, 169 of them.

I have put straight into Skip Technologies
that I am not, really interested in,

doesn't mean they're not useful, doesn't
mean you can't be interested in them.

These are personally things that I
do not give two hoots about because

they do not fit in to my stack or
my production or anything else.

doesn't mean they're not good, I have
already made my opinions on observability.

Thanos is a great project.

But I'm just not interested in that.

So it's in Skip, we've got
Holmes, GPT, just not interested

in bringing AI into my cluster.

I use AI everywhere else.

I do not want it in my cluster.

bank Vaults, Aqua
Project, ClusterNet, Dex.

I mean, I know some of these and some
of them are just not, if someone were to

click on it and say Go check, oh, OpenTofu
is going to move, that's, a no brainer.

Yeah,

because I was always Team Terraform.

motivations change right,
right, right, right.

So I get to update this and people
can follow along with my Stack It

doesn't mean anything in Skip is bad.

It just means I haven't got around to it.

It's not for me right now,
or I don't have a use case.

Yeah.

I wish I could be in a position where
I could add opinions on everything,

but again, that is a risky game.

This is a volatile ecosystem.

I've always felt, for me it's
always felt like a little bit

of like, stay in your lane.

there's different areas where I'm
like, you can have opinions, but they

would not be based on full production
experience and knowledge of running

clusters long term with these products.

I'd hate to have opinions when I
have a cursory knowledge or I've

done a sample install in my closet.

that's not enough for me to
feel like I have an opinion.

You also have graveyard on the other end.

Yeah.

So the graveyard, is projects that
are either abandoned, so people should

just get a warning not to use this.

I'm going to have to add
Terraform, CDK to it today.

or projects that have burned me.

and I feel bad by that, but
you know, sometimes a project

is just, shouldn't adopt it.

And I'm just graveyarding it so people
know, SaltStack, is not a dead project,

but it was bought by VMware then
Deprioritized and now that it's Broadcom.

I think you would be absolutely, crazy
to adopt SaltStack in 2026, but you

I was a fan a decade ago.

It was, my go to over
puppet, Ansible, everything.

I feel, very disappointed that
it's not my go to anymore.

And I felt bad about putting
cadence here because it literally

just got moved to the CNCF.

that worked on cadence went to Temporal
and Cadence was deprioritized by Uber

this was their last chance to try
and save the project by donating it.

I just don't see anyone actively
contributing or using it.

maybe I'm wrong, but I don't think
anyone picking that today over anything

else, is making a very naive mistake.

the graveyard is just there
as a, you know, make sure

your projects aren't in it.

And the nice thing is, you can come here,
and if we go to advanced mode, for the,

audience, it gives you search and the
ability to group and slice and dice.

you can just search, cadence won't
show up because it's graveyard.

is the NGINX Kubernetes project?

Controller

I don't think

in there.

The only one I've got is Emissary on Skip.

the ngres controller did
not make the cut, but I

guess I should add it to graveyard.

I was.

I do have to try and keep
on top of this now as well.

It's going to keep you sharp.

It's going to keep you sharp.

Is this a open source repo where
people can help you make corrections

or is it, I guess this isn't really
corrections, it's more like your opinion.

So how could other people
help you with that?

That's a great point.

So, we're on the GitHub repository
for the Rockwood Academy.

There is this wonderful content folder.

And you can go to technologies and
every technology is in here, If we just

grab, let's use salt, we'll just pick on
salt today, you can open the MDX file.

as you just marked it
with, a YAML front matter.

It's very easy for anyone
to come and change this.

If you need to update the, logo for
whatever reason, you can do that.

we have a matrix graveyard
here, and this is where I can

store my opinions under matrix.

But one of the things I would love to do
sooner rather than later, like before the

end of the year, is to just allow other
people to come and put their opinions

here so that you can actually click on a
face and it's like, what's david stack?

What's brett stack?

You know, what's X

y stack And just make
it this community driven

Stack

Yeah.

tool because my opinion, is
my opinion and it's my path.

And I hope other people find that useful.

we should have everyone's opinion.

Anyone who wants to contribute their
opinion should be able to do that.

the reason I didn't launch with
that is how do we allow this

to visualize if it grew beyond
four or five people's opinions?

Does it start to devalue the, matrix?

I need to work that out.

So there's a lot of things I would
love to do, but, getting more opinions

into it, and it is all open source.

People can come and contribute.

so lots of things on my mind
about how to push that forward.

Yeah, it feels like one of those things
where you could spend 40 hours a week,

on this, Cause I'm sitting here thinking,
yeah, you could probably make that it

shows opinions on the project I'm like,
I'm already trying to work out your

exact, your actual ui ux about that.

Let's get everybody's opinion.

Let's bring them on in.

But, but let's not elevate
everyone's opinion.

let's elevate certain voices.

Maybe if they've opinionated on five
projects, they somehow get to a different

tier there's a lot you could do with
that, we know about all these product

review websites on the internet that suck.

They all these, G2, there's all these
things that I find on the internet that

are like, well, we have user opinions
and we think this is better than that.

I might be comparing like, you
know, Excel to Google Sheets

or something, I don't know.

But those websites always felt
suspect to me and lacked genuine,

like, who is saying all this?

it does feel like a cool idea to
see opinions of different projects.

It would be interesting to even some of
those opinions could link, is it possible

to add other people's opinions that
they have put out publicly if someone's

wrote a blog post about it in the past
where they've had an opinion on, like

for me it's the classic Argo versus flux
discussion around which tool or both.

in that case it's I'm sure there's
a lot of opinions on both sides.

There's probably a very few amount of
people that have actually used both.

so that would be cool to see
their opinions or to have

links to their blog posts.

But yeah, this is the kind of thing where
there is no end, there's no end for you.

No, there definitely is not.

But, find ways to bring in more opinions.

It could even be that if you're just
on the Academy website, it pops up

and you pick between two technologies.

We can still collect information
that is valuable to people,

We can still build up
community scores for stuff.

lots of ideas and I'll push that forward.

Yeah.

there's 169 skips, there's
14 in the graveyard.

So, I've been working very
explicitly 14 of these technologies.

And then the rest of them are on
that gradient of be keeping an eye

on this one until you should start
actively learning this technology.

Yeah.

for the audio audience, can we run
real quick through the advocate?

The 14 and just list them for
the audience so they know of

these are, the heavy hitters.

everybody should be at least
a little curious about it.

Is that how you're
approaching it for teams?

It, doesn't mean they
have to implement it.

It's not like a foundational technology
that they would always, you're not

saying this is required, you're
just saying I'm a huge fan of this.

I'm a huge fan.

I would never build a production
platform without, these tools.

This is what

Oh, okay.

for joining us today.

we hope you have a great rest of your day.

We'll see you next time.

Take care.

you need a common format.

And cloud events is just so simple.

It's just a spec.

They have implementations
for serialization across

all different languages.

if you have any sort of semantic
data that travels over a network just

wrap it in a cloud events envelope.

It doesn't matter who you are.

everyone should have it in their
NPM, go dependencies, REST, Cargos,

It's one of those understated,
underutilized technologies that

saves a lot of downtime, bugs,
regressions, So yeah, amazing project.

I'm a huge fan of CUE.

I literally just published a
video on it two days ago for,

making GitHub Actions easier.

I use CUE across my entire monorepository.

I even released a tool called QNV,
QENV, whatever you want to call it.

last week to replace my, disparate
tools of Justfiles, Bazels, CI

CD pipelines, Dart env, configs.

All of it is now solved by one tool.

powered by CUE.

very, very cool.

Hello, and welcome back to the
DockerCon Monthly Dev Session.

My name is Matt Granitz and I'm the
Chief Editor of DockerCon, I'm joined

by Matt from Github, and Matt has been
really helping us out on this project.

How are you?

I am doing well.

I'm

The Zanzibar paper came out of
Google, nearly a decade ago.

SpiceDB is an implementation of that.

It will simplify your life if you
have permissions across the board.

don't reinvent the wheel here.

Use the tools that are out there.

I never have SSH on any machines,
so teleport is always right there.

I use Teleport for clustered.

I have done since 2019 when
we did the first episode.

again, you run a small binary
on every machine and cluster on

databases and it commoditizes access.

Yeah, I'm a fan of Teleport.

need to get a logo there.

Yeah,

I need to get the logo in for that.

Maybe someone on the teleport team will
contribute that to the repository for me.

They,

also gave away Yeti cups.

So, you know, they got the swagger
in their swag a little bit.

by the way, for those listening,
we've done cloud events, Q Flux

cd, spice db, teleport, and you're
breaking these out into areas you

mentioned that the, first three are
in app definition and development

rest

SPICE db, teleport, we've got
a few more to get through.

are any of these, it sounds like
these are all technologies that

aren't necessarily competitive.

So when you recommend or advocate
for something, are you picking

what you consider your favorite
or your winner in a category?

Or would you

see

there's two good service
meshes or whatever?

I'm, trying, again.

I'm trying not to be swissolent.

like there are more than two
good service meshes, but there's

only one I would adopt and that's

the

Okay.

I'll have under advocate.

I wouldn't have two there.

and funnily enough, we're now
into where Linkerd pops up, right?

Like that is my service mesh.

I use it everywhere.

it's simple enough that I don't
need to overcomplicate the setup.

It does what I want it to do, again,
without too much configuration.

It works by default.

It's lightweight, it's got a proxy
written in rust, like it ticks all the

boxes on the Dave Book of technology.

Like when I say, would I use this?

It's got all the ticks.

doesn't mean STO is not great.

Everyone uses STO successfully
in the enterprise.

Lots of banks use STO.

they have requirements that lean them
more towards it over something like

Linkerd, but to me, Linkerd is just
where we should be service mesh wise.

And it's that balance of power,
defaults and, doing the right

thing, like it's really important.

Obviously Kubernetes is here.

I've been advocating for
Kubernetes since, day one.

Will that be replaced at some point.

Yeah, probably.

But that day is not today.

These categories, by the
way, are from the landscape.

I didn't want to reinvent the wheel.

okay.

my own categorization of software,
which we can take a look at, but the

landscape is, I'm, trying to make
the landscape simpler, but again, I

haven't deviated too far from, that, we
then have cilium and EBPF, is my CNI.

I discourage people using any other
CNI because I feel like you're

just throwing out too much power.

EBPF is phenomenal.

XDP is phenomenal.

Hubble, who does not want Hubble in
their cluster, you're just missing

out on so much and making your life
easy EBPF, while not a product or a

project that one can just adopt as
something that solves the problem.

You have to learn it.

It's like a programming language.

I do advocate for EBPF.

It's a superpower to anyone that's
willing to invest the time to learn it.

then we get down to observability.

My stack is FluentBit.

I love FluentBit.

It's lightweight.

It's written.

C it's fantastic.

Just use it if it works.

There are times where
you can't use FluentBit.

use FluentD.

but FluentBit now I think has
enough parity that you're probably

safer using FluentBit anyway.

I've used it for a long time.

right on you

FluentBit, I've, yeah.

I felt like when it started getting
traction, I thought, well, I'm, going

to switch from FluentD, so, yeah.

Because who wants something
written in Ruby when they can

have something written in C

Yes.

Grafana needs no introduction.

Everyone probably has
Grafana on their stack.

OpenTelemetry is like CloudEvents,
it is now the universal language

for metrics, logs, traces, and
anything that comes in the future.

So just adopt it and simplify your life.

The last two are Spin
and WebAssembly and WASI.

I've been a big advocate
of WebAssembly for a while.

I do think it is the future.

we're not quite there yet, but I
will advocate the hell out of them

any opportunity I get because the
developer experience of building

applications that way is second to
none It is absolutely wonderful.

rust, I struggled with not
putting this in advocate.

I put it in learn because I do feel,
it's not that I should be telling

people to put rust in their production,
but everyone should learn rust.

It will teach you to
be a better developer.

This is just broken down
into a Trello board, Kanban.

you come along and pick the ones you want.

if you want advanced mode, you
can say, show me one of these.

I can just say, I only want to see.

What should I be learning?

you get this learn column, and these
are all the technologies that I think

will make you a better engineer,
developer, and operator if you spend

some time and invest in learning
them dagger for one as we click on

it.

Yeah, that's a good list for me.

because my students one of
the most common questions.

We get in our Discord server, we get
people in there every day joining

through the, discord discovery
when, they let us be in discovery.

because we get kicked out often
because of activity issues.

the IoT

we go into discovery, we get a
bunch of people and they're like,

onboarding is, low or whatever.

We're going to kick you out of discovery.

and then they're like, you meet
the requirements for discovery.

Project.

But anyway, the people
show up and I would say

an

I need a bot.

So when you say I'm learning DevOps,
I have bots that send you a whole

bunch of links in a dm. I, never set
it up, but I need to do that because

there's a lot of people, especially
in India, coming into tech and wanting

to go directly to DevOps activities
or platform engineering activities,

either at the beginning of their career
or very early in their career, you and

I have been doing this a long time.

these those terminology, those
concepts of the job role specifically

for that existed 15 years ago, but
it's not the same as it is now.

And the, what we call, what we think
of when we think you have a DevOps

role or a platform engineering role
today are different than they were.

And certainly the technology list is
completely different than 15 years ago.

there's very few places that are,
I think, good opinionated paths.

of course we've got roadmaps sh or
roadmap sh and they do a lot of good

work there, but even their simple
version of the learning path can be

quite intimidating, for new people.

these are the core foundations
for your overall IT learning.

When we get into platform engineering
or caring about clustering technologies.

There's nothing I can point people
to until today that says, okay,

I know there's a lot going on.

yes, you need to go learn Kubernetes.

But people often ask me about
what CNI, they should learn.

They often ask, about which logging
or monitoring tool they might learn.

I think those are maybe some
of the most common ones.

The Flux versus Argo
conversation comes up a lot.

I'm an Argo fan, but it's mostly
because that's what I've used.

I've used Flux very few times and it was
definitely before the last major version.

It was years ago.

So I don't even feel like I can have a
strong opinion on one versus the other.

I just know what I like.

So for the, is that kind of what
your audience this learning part?

I need a short link to send people
to the learning path because that's

something that we could put into our
Discord server for people to find quickly

for what they need to check up on.

I guess if they're brand new to all this,
they start with the advocate stuff, right?

I guess that's the challenge here
where do you meet your people?

you've given me an idea for.

How to frame the start of the matrix.

I mean, there should be
a, button, Why am I here?

I want to know

Hmm.

I want to know what
technologies are on the horizon.

I want to know what I put into production,
and I want to know what your stack is.

Those, are the different questions
that I'm trying to answer.

Just because these are my
opinions and this, is my stack.

Yeah.

you will actually get my blurb of
why I use this over the other tool.

I do try to explain why FluxCD is
the tool that I use for GitOps.

to make it a bit more fun, I've added a
spicy take to everything that I advocate.

I don't know if this applies to you as an
Argo user, but since you did say you use

Argo, you're getting the spicy take, and
for the audio audience, I'll read it out,

Yeah,

Flux CD is the grown up option, and most
teams don't realize it because they're

busy clicking around their dashboard.

sorry if I've offended you there at
the same time, but I, do Argo CD won

the GitOps wars because it had that UI
made the delivery of your image updates

very visual and people loved that.

I'm,

see

I am one of those people.

of flux.

I mean, yeah, you're not offending me.

And it's a great
conversation to have because,

We hope

that's stuck with me for years is that

I talk to teams that use both,

time.

that they use Flux to manage
the systems and they use Argo to

show the dashboard to the devs.

I've never implemented
both, at the same time.

But with Argo, I make the
dashboard read only so a lot of the

functionality is, just, you can't use
it because it creates bad behavior.

It's click ops.

And yet I still love the, I love
giving developers a way to see into my

clusters Giving them a regular dashboard
that would show a list of pods that

they maybe aren't, they're maybe more
concerned of, Hey, did the thing that

the job, that run, finish and is my app
actually being served on the website?

that was a common thing with a
couple of projects and they needed

something to tell them that.

And I was like, well, we got this
dashboard with Argo, I'll give it to you.

So I started giving it read only, and
then I got a lot of success with people.

So really my own path to liking
Argo was through trying it first.

Even though I was hoping at the beginning
I was like, well, flux seems cooler.

I'd like, I think I like this approach.

But you're right, I think the web
project, the web ui originally it was

shelved or it was a third party thing,
but it wasn't well kept or something.

I can't remember the
origin of the UI part.

Yeah,

I just remembered it wasn't,

one,

yeah.

the Weaveworks, enterprise one, was
available for people that did that.

it was really difficult to get that
web UI for Flux back in the day,

but now you're spoilt for choice
because our purpose built Web UIs I

know the Gimlet team, shipped one.

But not only that, headlamp has first
class support for Flux CD as well.

if you really want it, it's there.

But I'm of the opinion that, you have
to trust your GitOps pipeline and

you don't really need to watch it.

I tend not to have one.

when, you have two great projects,
or at least two popular projects, I,

sympathize with people because they
feel like they have to go super deep

in order to make an educated choice.

And who's got the time
to do that a dozen times.

So that's exactly what you just stated
earlier about why you created this people

shouldn't have to spend six months to
a year making decisions about the basic

technologies that go in their stack.

like you said earlier,
these are plumbing to us.

They are, essential components
of a modern IT stack.

And you need a lot of the things
you mentioned, We didn't really go

deep into logging or monitoring.

I think you did say Grafana, but you
didn't talk about Loki versus others.

And, I don't know if I saw
Prometheus on the list,

I wouldn't say that I specifically
always run one or the other.

when it comes to observability,
I tend to use a SAS product

rather than run on my own.

because I'm happy to, I'm happy to
pay the price there want more sleep

and I'm a team of one, so I don't,
really need to make that trade off.

If I were back in a company tomorrow
and responsible for a production

cluster, I'd probably be running Lokey.

I think it's such an amazing piece of
software and I'd probably running meur

because I know Grafana are gonna back that

Yeah.

Because InfluxDB have done
three major rewrites in.

Six years.

that's a scary prospect for
people putting it into a critical

part of their infrastructure.

I'm willing to take that risk
because I'm a big fan of InfluxDB.

I used to work there, which
is also why I'm a fan of it.

but I'd struggle to put it
in Advocate when's Influx 4

going to be written in Zig?

is that going to come
out in two years time?

I don't think they would ever do that.

I, think they have found the vision
that Paul Dex, the founder had.

I think he's now there and it's a safe
bet now, but would I bet my career on it?

I don't know.

So, to wrap this up because
we could talk for hours.

what are some key takeaways?

Do you have any key takeaways
for people this is for both new

people and experienced engineers.

If they're wanting to learn, if they're
looking for what they should learn next,

this will help them, I guess with that.

With that,

if you go to advanced mode and filter
for what's important to you, If you're

looking for application delivery and
want to know what to watch, learn,

explore and adopt, it's all right
there, plain black and white with my

feedback on as much things as I can.

And do you know my goal?

Is to add 10 pieces of feedback
to each project, or at least

every day, add 10 more pieces.

this is the V1 of the tech matrix.

yeah.

on major developments.

I, will do that as, things change
a usability point right now.

feedback is great.

Please hit out on socials if something
isn't right or I can make your life

easier or give you more opinions on stuff.

those questions that go to
your discord, send them my way.

I'll update the matrix to include that.

but yeah, know it's not going
to help for people that are,

listening, but use the advanced mode.

That's my one piece of advice to anyone,
no matter what journey you're at, go

to pipeline stage and just clear it
and select the one that makes sense.

if you just want to know what is
cool and upcoming, the watch pile,

there's 35 pieces of technology there.

Daytona is gaining a lot of ground
in ai, and development environments.

Start kicking the tires on that.

we've got Dalec with a

much for joining us today, Bye.

confirm that it's worth watching,
but I am still keeping an eye on it.

if we progress from watch to explore,
things that are safer bets might

be right for you and it may not be.

It just depends what you're trying to do.

One that I think is
really cool is Perseus.

This is like the Grafana replacement.

It's an amazing project,
but it's so early.

and it's not really got that multi
vendor contributions yet, They show me

that it's going to have that longevity.

So I think it's amazing.

I hope they do amazing things and
keep going, but it's very much

in that explorer phase right now.

I can't encourage people
to go and learn it.

another amazing project is Kairos
from the Spectral Cloud team.

this allows you to take any Linux
operating system and make it an immutable.

Cloud native Kubernetes running machine,
It's such a cool project and then of

course if you ditch the explore and
learn, these are technologies that I

think will pay dividends on your career,

So that question of, I'm new
to DevOps and I want to learn.

I hope I could guarantee that if you
pick anything on this list, it'll

be a net positive on your career.

Flatcar is an amazing piece of technology
that I think if you learn it and

learn how to use it, great things.

Same with Knative.

If you want serverless on Kubernetes,
there's probably no better choice

than, the Knative Project And Dagger,
an up and coming CICD pipeline,

that is all as code and TypeScript
and all of that really good stuff.

This isn't just new shiny tech.

Has anyone ever lost their
job from learning Postgres?

No chance, everybody
should learn Postgres.

it's going to make you a better DBA.

It's going to make you
design better schemas.

It's going to make you think about
indices and optimizations and

performance at the database layer.

You can even go down a rabbit hole of
learning P-L-S-Q-L, and that will still

pay you dividends later on in your career.

The way.

of software.

I love that you're like, I, still
have the, 'cause I love that.

you also have the Kubernetes
project listed there, which I

think is, an assumed, an assumed.

Yes.

I?

I want to do a little plug
for my friend Aaron Francis.

He just launched a Postgres
course called Mastering Postgres.

he has built a bunch of database
courses at databaseschool.

com.

That's an unpaid ad for my friend Aaron.

He's a wonderful guy and, he's
making really good content

and he's making me jealous.

because he rented an apartment,
built a whole studio in it, put an

actual blackboard in there, like a,
teacher's classroom and is killing

it in the content game right now.

think,

courses with a lot of technical

little

deep insight.

I, feel like the same way.

actually

it's like.

the

To me,

you need

networking, getting a Cisco
cert early in my career, in the

nineties, was pivotal foundational.

I still use that knowledge every day.

of that

I don't necessarily have to break down
my, subnets into binary a whole lot

anymore, but understanding that stuff
has always helped me in every level

above that in the technology stack.

And that's to me, like TCP IP is one
of those things that it's almost feels

like people learn last now because the
cloud just sort of glosses over it.

But databases are the same thing.

you and I could probably walk
into a typical cloud native shop

of a small team that's maybe in a
little startup I'm presuming that

you've never been a full time DBA.

I would never have called myself that,
but I've managed lots of databases over

the years and I bet you I could still.

see

after I first ranted about
indexes, I could probably still

walk into a team today and they're
having performance problems.

And I'm like, yeah,
you're not doing indexes.

let's do some indexing.

to me that's like a foundational
understanding of that anyone

should have a database.

It's Like even if you're just a, DevOps
engineer a full time ops engineer running

Kubernetes clusters, like, you should
probably understand the basics of how

SQL works and how SQL databases have
logs and how they flush some of these

sort of basic technologies that we
can't avoid even if you outsource it.

So it's cool that you had that stuff
in there because that's key stuff too.

I mean, it's like you're, you're getting
right on the edge of almost being.

it's you're the cloud native stack.

So I go to DevOps sh for my, which
programming language should I

learn for DevOps, which, you know,
it's got the typical Ansible,

Terraform, blah, blah, blah, Docker,
Kubernetes, learn all these things.

learn a cloud, get a
cert in DevOps somewhere.

And then now it's like, okay, now you
need to be actually productive and make

decisions on how you're going to run
your web servers and your databases.

And that's maybe where the cloud native,
like we jump over to David's website

and that's where it starts off with.

And I feel like there's
a continuity there.

We could sell that.

I hope there is value to
everyone who is trying to learn.

the learning doesn't stop.

This isn't people, you know, when I
say people trying to learn, it's not

people that have got 10 days experience.

You know, even people that have been
doing this for 20 years, like us.

is there ever a day or a week that
goes by where you're not learning?

and the Postgres thing, right?

what you were just talking about
indexes and all of that, You've

got to get the basics right.

You've got to master the
basics 10, 000 times.

like the old, I can't remember his name.

The martial artist that died

Bruce

Bruce Lee.

he said he didn't fear someone who
practiced 10, 000 kicks one time.

He feared a person who practiced
one kick 10, 000 times it's true in

development and engineering, you've
got to get those basics down and then

that gives you the, ability to do it.

the freedom to go and explore
all these other technologies.

The T shaped developer is
a very important paradigm.

that changes when we bring AI into it.

is my matrix already useless?

Because people can just sit
down with Gemini and say,

what should I learn this week?

It's like, I don't know.

But I still think there's a lot of value
and, you know, the human touch and our

opinions and sharing that with people.

And this is my way to try and
ease that burden for people.

Nice.

Well, thanks for creating this yet.

Another thing for the community, I feel
like your list is ever growing of, of

things you've, you're giving away in terms
of content and advice and edutainment,

which I also have used that word before.

So that's good that you think that, 'cause
I think that is key is a lot of this

stuff can, can be dry and dull and we've
already got that in a thousand conference

talks every, well, every three months now.

I guess it was definitely more
than a few thousand a year of

conference talks on YouTube.

So we don't need more
technical content necessarily.

We need something that's a
little better, higher produced.

And you are one of those gentlemen
that I. I like to watch your videos.

I watched your cue video,
from a couple of days ago.

last night actually, I think I,

So,

up in my feed.

be

you're somehow YouTube knows
that if it talks about cloud

native, it needs to show up

later.

next to late night comedy
and stand up comedy.

Like those are my three things on YouTube.

They're all nice and mixed together.

Oh, and, EVs, anything ev and it throws
you and Victor Farsik and some other,

Sid and some other good YouTubers.

it throws them all in there in the mix.

great content.

Yeah.

So thanks for doing this and I'm looking
forward to getting this podcast out.

And you can find more of this at, rawcode.

Academy and all the links are
gonna be in the show notes.

David, where can people find you?

I know you've also got a Discord server.

do you want to tell the people
where they can find you at?

Yeah.

you can go to Discord, GG slash rawcode.

I'll go to Discord server, I am
rawcode on every reputable social site.

and, you know, even know what's there.

the.

way the world is right now.

I find myself on LinkedIn a whole
lot more, which I never thought would

happen, but I am active on LinkedIn
sharing my opinions there too.

So I try to be everywhere.

and I'm always happy to
sit down with people.

I have a very public office hours link.

I'll make sure Bret has
that for the show notes too.

if you ever just want 20 minutes
to sit down and talk tech,

it's there for the community.

I have nothing but time.

Yeah, that's great.

That's perfect.

this is the exact DM I sent to a group
of, cloud native friends recently.

Random thought.

I get more value from scrolling
LinkedIn than any other platform.

I actually enjoy it.

I think 90% percent of
what I see is good quality.

What a strange world we live in.

Oh, never thought I
would see the day either.

And here I am.

I'm not only consuming LinkedIn,
I'm now posting to LinkedIn,

which is a very new thing for me.

But I do think it's given people a voice
and I want more voices in the community.

I want to get more opinions,
more feedback, all of that.

And it is one of the
better places right now.

Definitely.

yeah, Low, drama, high quality.

that's at least my feed.

I'm very lucky in there,
at least right now.

I, didn't do a lot of work to get
it there, so I assume that that's, a

experience that others might have in tech.

I know Blue sky is a thing.

I'm a big fan of Blue Sky, but,
the community at least the cloud

native community, And the, the
container and cloud community

feels more active on LinkedIn than
on, BlueSky, at least currently.

thanks for being here and I'm
sure we'll do this again and

see each other at KubeCon,

bye.

You can sign up for my newsletter at bret.

News to get this and more
cloud native content.