Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
Clough cause media presents from the
0:02
massive studios the valley North Carolina.
0:04
this is the Clouds cast with
0:06
our and Dell and Brian gracefully
0:08
bringing you a best of cloud
0:10
computing from around the world. Good.
0:13
Morning you didn't rub your and will come back to the. Cloud
0:15
Guest were coming to live former massive grow
0:17
Guess studios are in Raleigh, North Carolina.
0:20
And. It is Aaron for a quick
0:22
intro this week. We've. Been gathering
0:24
news all month for next week's clone news
0:27
show. It's. Been an interesting
0:29
to say the least month in the
0:31
emerging deck space and you'll want to
0:33
be on the lookout for that next
0:35
Wednesday. This. Week though, we're
0:37
talking about visualization of data and diving
0:39
into observe ability I've been wanting to
0:41
talk to grow fond of for a
0:44
while now and I'm glad we're able
0:46
to make that happen and we're going
0:48
to jump right into that right after
0:50
this break. Are. You looking to stay
0:52
ahead of the curve in the latest trends
0:54
and enterprise tech, look no further than Breaking
0:56
Analysis podcast with Dave Valenti. This data driven
0:58
program dives into the most important topics facing
1:00
the enterprise tech industry today. With. The
1:02
data first approach that leverages eg. hours from
1:05
now on, surveys of I T decision makers
1:07
and insight from the cube community. Breaking.
1:09
Analysis delivers indepth research on the
1:11
most important topics facing technologist and
1:13
I T buyers. Whether. You're a
1:15
business leader, an I T, professional investor, or
1:17
just an avid follower of the industry. This
1:19
podcast is a must listen! Just. Search
1:21
Breaking Analysis Podcast wherever you get your podcast
1:23
and tuning today to stay ahead of the
1:25
game and enterprise Tech. Transform.
1:30
Your business with Microsoft Azure The one stop
1:32
shop for everything cloud in a I. Would.
1:34
Azure migrated modernize. Move your existing
1:37
workloads to Azure with ease. With.
1:39
Azure Innovate, build new solutions with
1:41
focused and deep engagement. Their.
1:43
Core offerings provide comprehensive resources,
1:45
expert assistance, and cost optimization
1:48
tools. Accelerate your cloud journey
1:50
and stay ahead of the competition with Microsoft
1:52
Azure. Visit aka.ms
1:54
slash Azure Hero Offerings
1:56
to learn more. That.
1:58
That's aka.ms slash Azure
2:01
Hero Offerings to learn more.
2:06
And we're back and Aaron
2:08
it's good to be back doing this with you.
2:10
I am I am finally back after the the
2:12
much Bali who'd trip around the world. How
2:16
are you man? How have you been? I'm
2:18
good. I'm good. I more than anything I
2:20
want to know like what time zone does
2:22
your body think it is now that you're
2:24
back? How long what's the over under and
2:26
how long it's going to take for you to adjust? Yeah,
2:28
no idea. No idea. I'm in for a month now I
2:30
wake up every day between two and three in the morning
2:32
and I go to bed at eight
2:34
o'clock at night in whatever country I'm in. So
2:36
I have I have no idea. It'll probably take
2:38
a week or so to kind of get back
2:40
to normal but could be back in it back
2:42
in North Carolina back in the regular massive studios
2:44
and you know speaking of sort of you know
2:47
big things and big visibility about stuff which you
2:49
know I've been seeing the world. We
2:53
have not had a chance in a little while
2:55
to kind of talk about the thing that that
2:57
sort of sees everything and sort of looks at
2:59
things globally which is observability and today we're going
3:01
to kind of dive in and be like what's
3:03
new with observability? What are people learning? Is there
3:05
as they're deploying it all over the place and
3:08
why don't you go ahead and introduce our guest?
3:10
Yeah, absolutely. And so real quick
3:12
this is interesting too because this
3:15
actually comes out of a conversation
3:18
our guest and I we had at all
3:20
things open this year in Raleigh. So
3:23
I stopped by the booth said hey and
3:26
one conversation led to another and here we
3:28
are with doing the podcast and
3:30
so what we have this week is we're
3:32
going to be talking about obviously observability and
3:34
visualization of data in general and so we
3:37
have Ronald McCollum senior manager solutions engineering at Creative
3:39
Final Labs. So
3:42
Ronald welcome to the show. Oh,
3:44
thank you Aaron. I'm really happy to be here.
3:46
Thank you for having me on. Yes, absolutely. Before
3:49
we dive into today's discussion why don't you give
3:51
everyone a Brief introduction to your
3:53
background please? Yeah, absolutely. So, yeah, since
3:55
we met at all Things Open I
3:57
think it kind of follows that I've
3:59
been an Open. and for escape from
4:01
way were back I got started in
4:03
the linux and open source community is
4:06
back in the mid nineties of he
4:08
hopes in high school went from there,
4:10
became a programmer for a bit but
4:12
I i really always enjoyed the intersection
4:14
of a software and hardware. I like
4:16
to play with an infrastructure elected he
4:18
to hold a screwdriver occasionally so I
4:20
ended up in the monitoring space keeping
4:22
the lights on and Adidas other up
4:24
and after that of have worked at
4:26
a number of companies including Canonical which
4:28
is the company behind. Have been to
4:31
Linux where I built up the original
4:33
hardware certification program there so I got
4:35
the scratch that itch of combining hardware
4:37
and software again. Ah, and really from
4:40
there have been in a bunch of
4:42
small startups which included your photo labs
4:44
when I joined about five years ago
4:47
although it is way way better now
4:49
I and I managed to team solution
4:51
engineers. They're helping folks filled out observe
4:53
ability practices with the group. Honest. Testing
4:56
and and I'll say this real quick round
4:59
to I'm. So. We last talked
5:01
to grow fonder. Gosh, we're actually couple
5:03
times a twenty nineteen and and twenty
5:05
twenty and in links in the sooners
5:07
by the way. see if anybody wants
5:09
to go back and times observe ability
5:11
continues to be a really hot topic
5:13
on. In this combination of
5:16
roof observer ability and open source,
5:18
how are you seeing the open
5:20
source community and open source tools?
5:22
And how's that involved in this
5:24
space? Because it has been quite
5:26
a few years, Definitely, and I
5:28
really honestly believe that open source
5:30
is the natural and point for
5:33
most software. but definitely observe ability
5:35
and especially things like data collection.
5:37
So I kind of think of
5:39
it like plumbing where it's something
5:41
that you really don't normally think
5:43
about, right? It's. There it's in the
5:45
background and as long as everything's going okay,
5:48
you just ignore it. But. when
5:50
something breaks you suddenly care a lot
5:52
and you want to fix things very
5:54
very quickly so just like plumbing really
5:56
you wanna be able to both get
5:59
in there yourself and fix things if
6:01
you can, but also have a ton
6:03
of people who are already familiar with
6:05
the things that you're using can jump
6:07
in really quickly and get things fixed
6:09
and back up and working. So I
6:12
think about open source in that way
6:14
that if you are relying on proprietary
6:16
tooling for your observability system, you're
6:18
both really reducing the size of the talent
6:20
pool that can work on those tools, but
6:22
you're also locking yourself into one vendor. So
6:24
if things change out from under you, you
6:26
kind of have to go along with it
6:28
or do a very expensive lift and shift
6:31
off of that. So I think
6:33
as a result of that, we've
6:35
seen the particularly the data collection side
6:37
of things become commoditized and trend toward
6:40
open source, right? Prometheus has pretty
6:42
much one on the metric side. I think
6:44
at this point, we're seeing a ton of
6:47
momentum behind things like open telemetry for tracing
6:49
and probably eventually logging. So I
6:51
think that the real value of these things
6:53
is what you do with the data once
6:55
you collect it, rather than focusing
6:57
on kind of building out your own
7:00
data collection system, your own observability
7:02
telemetry system. Yeah,
7:04
no, it makes makes a ton of sense.
7:06
We've had Grafana on the show a few
7:08
times, we've tried to do our best to
7:10
follow kind of both observability
7:13
and dashboards. You
7:15
know, we might be naive in this, but
7:17
I think we oftentimes
7:19
hear Grafana combined with other
7:21
things. Grafana with Prometheus,
7:23
Grafana with some
7:26
set of sort of logging slash observability.
7:29
Where do you see sort of the
7:31
evolution of maybe both the role of
7:34
Grafana, but also maybe more
7:36
importantly, kind of the role of weird
7:39
dashboarding and visualization fits? Are
7:41
you seeing it mostly adopted by
7:44
infrastructure and ops teams? Is it
7:46
security teams? Is it individualized dashboards
7:48
for certain developers and certain applications?
7:50
Like how has it sort of evolved in
7:54
that sense of you know, both the
7:56
breadth of what Grafana does, as Well as
7:58
you know, whose. Using the
8:00
visualization, how are they best using it. Yeah,
8:03
I think it's a great question if
8:05
you have to look historically I think
8:08
you can really plot a line at
8:10
and feel excuse upon from people doing
8:12
things very ad hoc you know, starting
8:14
with logs In really at the back
8:16
to the very earliest days of computing.
8:20
and then really evolving from there to
8:22
add new ways of working with a
8:24
that thinking about data. So you kind
8:26
of move from this idea of individual
8:28
logs telling you about. What's. Going
8:30
on in your system to metrics that
8:32
can tell you more about things and
8:34
aggregate And as we've moved into the
8:36
cloud we see things like treating tools
8:39
that we can see how requests will
8:41
pass through very complex environments and and
8:43
touch multiple different services. So I think
8:45
if you you look at a visualization
8:47
side of things, it's really evolved in
8:49
sort of the same way. We we
8:51
started with just looking at raw logs,
8:53
having a developer kind of look at
8:55
the state of an environment, the state
8:58
of a service and figure out what's.
9:00
Happening to as those systems grew beyond
9:02
what an individual developer could look at.
9:04
having to figure out how to do
9:06
things were like our D tool and
9:08
Nog, ios and and other similar tools
9:10
that were the sort of early entrance
9:12
into that visualization and and graphing layer
9:15
of things. And then again as things
9:17
started to get even more complex you
9:19
see tools of all to be less
9:21
about directly visualization at the visualization still
9:23
important as the wanna know what's going
9:25
on but I start to think about
9:28
things like what can I automate what.
9:30
Can I have twitter and alert and direct
9:32
me to what I need to see? rather
9:34
than having somebody sitting in an operations center
9:36
all day just looking at screams and waiting
9:38
for something to go wrong? And
9:40
I think that. Really?
9:43
Ties into the the other side of
9:45
the question about who uses these tools
9:47
and how they use them. I think
9:49
it really depends on the audience that
9:51
you're talking to. So like if you're
9:53
talking to business, he asserts that the
9:55
folks that are running the actual business
9:57
itself and selling things to customers big.
10:00
Here a lot about things like return on investment.
10:02
They want to know that if they are. Investing.
10:05
In an observer delivery system, Investing in these
10:07
tools that is actually going to get the
10:09
more money. In the end, they don't want
10:11
to spend money for no good reason and
10:13
so they're going to care about things like
10:15
response rates or latency. If you've got an
10:17
E commerce solution, you're going to think about
10:19
the things that impact your users because if
10:21
they have a bad experience, they're not gonna
10:23
spend money on you. But.
10:26
If we think about this from an engineering
10:28
perspective, I think people think about. The.
10:30
Things that they want to observe very
10:32
differently so they focus on of things
10:35
like it's does your latency but in
10:37
with an individual service rather them latency
10:39
in at a summer transaction for example.
10:41
And their the folks that are gonna
10:44
care very deeply about things like infrastructure
10:46
and do I have enough sleep you
10:48
desk memory or cloud environments to support
10:51
the services that I have now and
10:53
can I continue to do that as
10:55
these things grow on something like Black
10:58
Friday. So I really do think that
11:00
there's. A couple of different audiences here
11:02
meet the care about different things, but
11:04
they're up closely enough. Related to that.
11:06
I think you can unify those two
11:08
worlds and and give everybody one view
11:10
of what's happening in their environment. And.
11:13
And rama when I think about
11:16
that in the context of of
11:18
your answer their he reminds me
11:20
of the early days of have
11:22
developed at times of I'm there
11:24
are some. Political. Considerations:
11:26
There's some cultural and cultural changes
11:29
considerations because this does impact anything
11:31
from you know, the business side
11:33
of the house to the the
11:35
asari side of the house, to
11:38
development. Ah how.
11:41
Do. You approach something when you're talking
11:43
to customers or or folks out there.
11:46
When it comes to bridging
11:48
those gaps, I'm or having
11:50
to build those alliances or
11:52
do sometimes in large scale
11:55
change. In Operations On that
11:57
seems like a delicate balance and
11:59
or maybe a recipe for disaster.
12:01
Sometimes out of the your thoughts
12:03
on that. Yeah, I have absolutely
12:05
seen people do this incredibly well,
12:07
and I've seen people do it
12:10
incredibly badly. And I think there
12:12
are. That.
12:14
There are lessons to be learned. their I'll I'll say it that
12:16
way. I really think
12:18
that the folks that are the
12:20
most successful in both building observed
12:22
ability, practice, but also in building
12:24
really reliable systems are the ones
12:27
that do it solas. Stickley that
12:29
it's really a top down. How
12:32
a drive from the sea level are
12:34
all the way down through the organization
12:37
that we're going to focus on quality.
12:39
We're going to focus on user experience
12:41
and as a result of that we
12:43
need to know what's happening in are
12:45
in agreement at all times. So the
12:47
organizations that I see that a really
12:49
successful in that way or the ones
12:51
that are thinking about things as a
12:53
serviceable agreements service level objectives right? They
12:56
have essays where their customers if I
12:58
don't. Perform. A certain
13:00
number of transactions in a in a minute
13:02
where I don't ship so many widgets and
13:04
our than I have to refund money to
13:06
somebody. So of course I'm gonna care very
13:09
deeply about making sure that I am getting
13:11
injured or targets and as a result I
13:13
can drive those as hallows, the service level
13:15
objectives internally to measure those things and to
13:17
know what I need to set to be
13:19
able to to meet those targets. And.
13:22
The reason I think that the
13:24
folks that use this approach of
13:26
the most successful are because that
13:28
Esa low focus. Actually let's you
13:30
decide where you should put your
13:33
engineering efforts right now. If.
13:35
you've got him as hello you can effectively
13:37
think about that as an error budget it's
13:39
how often can i miss my target and
13:41
still not blow my i select so if
13:43
i'm at one hundred percent i haven't missed
13:46
any targets for the month so far i've
13:48
got a lot of basically budget to work
13:50
with i can be really aggressive at building
13:52
new features rolling out experiments testing things out
13:54
and if a sale okay i can just
13:57
walk it back on still within those agreements
13:59
whereas on flip side, if I'm close
14:01
to violating those SLOs, I know I
14:03
need to focus very hard on stability,
14:05
on keeping things up and running, making
14:08
things more repeatable and calming down that
14:10
rate of change. So
14:12
yeah, I think if you if you
14:14
have that focus, that's ideal. But
14:16
I would say that I have seen people
14:18
be successful with a ground up approach in
14:21
building that kind of focus as well. You
14:23
can actually drive that from an individual
14:26
engineering level too. And I
14:28
think the way you do that is
14:30
you just start making things publicly available,
14:32
you collect information, you put some telemetry
14:34
and monitoring in on whatever you can,
14:36
and start sending that out, send an
14:38
email out once, excuse me, send
14:40
an email out once a week to all
14:43
of the managers in your organization that shows the
14:45
SLOs that you've put in place. And
14:48
when that happens, what will happen is those
14:50
managers will start to say, hey, why aren't
14:52
my services in here, I want
14:54
to show off how good I'm doing. So they'll
14:57
start to add those SLOs, they'll come to you
14:59
and say, how can I get included in this
15:01
report. And you run that for a
15:03
few months, you start to get a real groundswell of
15:06
people engaging with that kind of process
15:09
and starting to build the ideas of
15:11
running a real SLO focused organization. It's
15:13
honestly really, really cool to see when
15:15
that happens. Yeah. You
15:19
know, I tend to think about because
15:21
the space is interesting, it oftentimes, and
15:24
I kind of feel like I'm repeating that the previous question,
15:26
but I feel like we often see a lot
15:29
of tools get involved, right? And it tends
15:31
to break down to be, you know,
15:34
identify a problem or identify a situation,
15:37
try and resolve it if it's problematic
15:39
enough, you know, it's in the red,
15:41
it's above a threshold. And then oftentimes,
15:43
there's maybe a third phase of it
15:45
that might be something
15:47
that's looking at things historically, trying to maybe
15:49
do some predictive stuff for, you
15:51
know, kind of get in front of it. Where
15:54
do you I guess, two parts of the question,
15:56
where do You see Grafana
15:58
today. sort of sort of. The
16:00
the best at a been part of
16:02
that. The. String
16:04
of of trying to make life better
16:06
for operations and for applications and have
16:08
you as grow fond of tried to
16:10
you know kind of expand where their
16:12
role sets or do you guys feel
16:14
pretty comfortable with? You know the capabilities
16:16
you provide today and in where you
16:18
integrate with some of the other tools.
16:20
whether it's her you know, a logging
16:22
tool or some other type of thing.
16:26
Yeah, I think historically Griffon
16:28
I was very focused on.
16:30
The identification stage. it's is what is
16:33
happening in my environment right now and
16:35
that is because Grow Fonder really started
16:37
as a visualization tool, something that you
16:39
can plug on top of multiple different
16:42
sources of data and is visualize the
16:44
Molinar common way. And I think that's.
16:47
Often. Times what people think of when
16:49
they think a grip on up. But
16:51
really, over the last five years or
16:53
so, the platform itself has evolved, so
16:55
there's a lot more to it than
16:57
just visualization. That. Part still there,
17:00
but we've expanded the stacked much
17:02
more into covering. Really? Both
17:04
sides of the identification
17:06
and and remediation. Spectrum.
17:10
So. That We've got
17:12
tools now to do things like manage
17:14
alerting and notifications and escalation policies so
17:16
we can deal with events as they've
17:18
happened. Get the right people involved and
17:20
and surface that data up. Ah, as
17:22
well as things like back and there's
17:24
a skeletal previous back and the scale
17:26
of a logging backend of treason. Back
17:28
him than so forth for Griffon Out
17:30
that is just part of the native
17:32
stack there, so they're available to you
17:34
if you want to use them. Of
17:36
course, you can still plugin any data
17:38
source that he had before, but. I.
17:41
Actually, think that if you think about
17:43
this on a spectrum of like identification
17:45
and resolution of problems, it sells observe
17:47
ability a little bit short as a
17:49
practice because that is thinking about things
17:51
as how you react to issues. How
17:53
do I hello my notified about something
17:55
going wrong and how do I then
17:58
fix it. Were. if
18:00
you're doing observability the right way, according
18:02
to me, you
18:04
really want to be able to
18:06
understand what's happening in your environment
18:08
before an incident occurs. You want
18:10
to head off those problems before
18:12
they impact your users. So you
18:14
want to be able to do
18:16
things like load testing and scale
18:18
testing and do continuous profiling to
18:20
identify hotspots in your code. And
18:22
all of that's still related to
18:24
observability. I'm still collecting information about
18:26
my environment and about how my
18:28
services behave. But ideally, I'm doing
18:30
that upfront so that I can
18:32
identify those issues before they impact
18:34
users. So that's a real
18:36
big area of focus for Grafana right
18:39
now is building out that kind of
18:41
tooling, the ability to do those scale
18:43
tests and performance tests and continuous profiling
18:46
on your code before it actually goes
18:48
into production. That makes
18:50
sense. That makes sense. And so, Ronald, I'm
18:53
going to ask probably
18:55
the question that gets asked a good bit
18:58
these days. The
19:00
AI question. How
19:03
has AI either already changed
19:05
or in your opinion will
19:07
change both observability
19:09
and visualization? AI
19:14
is a big buzzword for sure. I
19:16
hear people talk about it almost on
19:18
a daily basis. And part
19:20
of the problem is I think every
19:22
single person that talks about AI has
19:24
a different idea about what AI actually
19:26
means. So it's very much the wild
19:28
west right now. But one
19:31
thing I have noticed is the biggest
19:33
divide seems to be between the suits
19:35
and the geeks on this one. So
19:37
like when I talk to business execs,
19:40
there's a big focus on things like
19:42
AI ops. They want automation,
19:44
they want auto remediation, they want
19:46
something that they can just drop
19:48
into their environment and have it
19:50
solve all of their problems and do
19:52
that automatically. But If I
19:54
talk to engineers, the practitioners who
19:57
are building the systems that they're
19:59
monitoring. The observe ability systems.
20:01
The things that they're interested in
20:03
are much more about surfacing the
20:05
right information and filtering out of
20:08
the noise so that they can
20:10
get to the right place and
20:12
fix the issue themselves. And honestly,
20:14
I think that ladder approach is
20:17
probably more real, at least for
20:19
the next few years. Ah, I
20:21
don't think it's impossible to automate
20:23
remediation to automate some parts of
20:26
incident response, but. Anything that
20:28
you can trivially automate is also
20:30
a pretty trivial to solve and
20:33
therefore to prevent selects. You really
20:35
do need a human level intelligence
20:37
involved in incident response and in
20:40
dealing with a of it incidences.
20:42
they occur in your environment. So
20:45
I think that. The companies that
20:47
are going to be the most successful over
20:49
the next four or five years in the
20:51
Ai space are the ones that are working
20:53
to enhance humans rather than replace them to
20:55
the ones that are are making humans more
20:57
effective. I I want
21:00
to sort of a follow up on. That
21:02
is do you get the sense? ah enough,
21:04
you're just put a stake in the ground
21:06
today. It like you said, you've got. You've
21:09
got mansion folks who are like look, I'm.
21:12
You know I just want the system to
21:14
fix itself. I wanted to be self healing
21:16
I just wanna drop it in in it's
21:18
in like a get less complaints from your
21:20
users in your constituents since and typically. Your
21:23
even just like automated systems, you have a
21:25
tendency of of the operations teams to sort
21:27
of be like. Well. They
21:30
don't know as much as we do. Not
21:32
really sure I trust the thing to do
21:34
stuff with the what do you think is
21:37
closer to be think that the technology is
21:39
is fairly close to being able to do
21:41
those things and. It's a matter of
21:43
of kind of. Convincing the
21:45
operation seems to be like it's okay. You
21:47
should let it do. These things: are you
21:49
still think we're. Quite. A ways
21:52
away from in of the gap between
21:54
sort of that that really smart operations
21:56
person that has to jump in there
21:58
and the system being. And
22:00
an equal level to where you know
22:02
that smart operations per second. In. A
22:04
quote unquote. Gonna go work on other
22:06
high value things. He? yeah. I
22:09
think there's some of of each to
22:11
be totally honest states. I do talk
22:14
to practitioners, the engineers who are concerned
22:16
about Ai taking a job. They have
22:18
been doing this for thirty years. I
22:20
I don't want to change and all
22:22
on or replace it, but I don't
22:24
think it's all that I think if
22:26
you think about. Well. Yeah, there's
22:28
the old joke about Ai that anything
22:30
that we can't yet do is artificial
22:32
intelligence, and as soon as we figure
22:34
out how to do it, it's just
22:36
automation. Ah, I think you see a
22:38
lot of that in observed ability and
22:40
in. Particularly. Cloud, skill
22:42
or infrastructure and I am as right
22:44
like communities as a great example There
22:46
to that it is is automation. It
22:48
is are all about self healing. It's
22:50
about being able to the to describe
22:52
how you want something to deploy and
22:54
then we'll let the system itself scale
22:56
it, upscale it down, kill it if
22:58
it die, if it it's into a
23:01
bad state restart things like that. So
23:03
those kind of automation I think are
23:05
going to continue to grow on it.
23:07
I think they're going to continue to
23:09
be a big part of the industry.
23:12
I think we're. We're
23:14
people might get off track and I
23:16
think we're some of the complexity isn't
23:19
always clear to the the business level
23:21
folks that are talking about things like
23:23
a i asked is that. You're
23:26
dealing with. Often.
23:28
Times when you're dealing with incidents that
23:30
are. Novel. The things that you
23:32
didn't expect. You didn't know how to look
23:34
for them up front. And so to expect.
23:37
An Ai system to be able to figure something
23:39
out that you didn't know about ahead of time
23:41
and didn't know how to even. Detect
23:44
let alone prevent. It's gonna be very
23:46
difficult to build a system that can
23:48
truly automate not only the detection side
23:50
of that, but the remediation side of
23:52
that. So I I think unless Chat
23:54
U P Five takes over the world
23:57
and and is truly sanction. ah I
23:59
don't see. That becoming were
24:01
that? that? The. State
24:03
of the industry ends up. The
24:06
mess up. My that's that's really
24:08
helpful Thank you on and so
24:10
our all out by you are
24:12
a quick of the final question
24:14
hopefully this is a softball question
24:16
are because you've. Literally.
24:18
Written a book on grow fauna on suffered
24:20
the hoses and we're interested it out there.
24:23
it's how do you recommend? They get started
24:25
with grow fonder know by the way. linked
24:27
to the book in the show notes. but
24:29
glad that a soft spot as an awesome
24:32
plug been que eran. I appreciate that. Yeah,
24:34
I tried to distill as much as I
24:36
could end of my books. And for folks
24:39
that like to learn things up front, you
24:41
know, really dig in before they get started.
24:43
A I think it's a great way to
24:45
a great place to start. I'm definitely. Going
24:48
to recommend my book. But honestly I'd also
24:50
say that Griffon A itself is pretty easy
24:52
to get into. You can go download it
24:54
for free, you can deploy it and on
24:56
your laptop or a raspberry pi are in
24:59
a container in the cloud. It it really
25:01
is very easy to get started there. you
25:03
can even use it. There's a freak tear
25:05
on grip on a cloud so I'm actually
25:08
more of a get my hands dirty kind
25:10
of person. So that's where I would start
25:12
and I'm always going to recommend the people.
25:14
Just in a kick the tires start playing
25:16
with it. I think or
25:18
Fauna is probably the easiest part of an
25:21
observer abilities stack to get started with. Because
25:23
it is pretty lightweight, it's pretty easy to
25:25
crank up and and start running. The hard
25:27
part? the real work is actually and thinking
25:29
about your observed ability Strategy: If you're always
25:31
gonna have to figure out what is it
25:33
that I care about, What signals do I
25:35
want to pay attention to them? What do
25:37
I want to do with those signals? So
25:39
there's no, again, no Ai that can really
25:41
do that for you? Yet you're always gonna
25:43
have to put in some of that work.
25:46
But have you not? Griffon itself as. A
25:48
very easy place to visualize the results of
25:50
that. and arts
25:52
that's very cool it's dwell on it it's also good
25:54
to know that it's it's not it's not so hard
25:56
that issue either you you potentially could have called the
25:59
book out you know I'm a fan of the hard
26:01
way, a la Kelsey Hightower back
26:03
in the Kubernetes days. I'm
26:06
kind of convinced that every way is the hard way
26:08
with Kubernetes, but that's just because I've been doing Docker
26:10
and bare metal for too long. Nice.
26:14
Well, listen, we want
26:16
to thank you so much for the time today. It's
26:18
been good to both, you know, kind of get
26:20
an update. You know, for us, it's
26:22
been a couple of years to get an update, but also
26:24
really kind of dig into, you know,
26:26
kind of your specialty, which is not just the
26:28
technology, but how is it used, what are the
26:30
best ways to use it? And so thank you
26:33
so much for, you know, allowing us to dig
26:35
into that. We really, really appreciate that today. Any
26:38
last, you know, one last thing, I guess, if
26:40
people want to follow up other than the book,
26:43
you know, kind of best ways to maybe engage you,
26:46
engage your team, engage, you know, getting started
26:48
with the technology that you might want to
26:50
throw out there. Yeah, absolutely.
26:52
People can always email me
26:54
at ronald.grafana.com, but feel
26:57
free to contact us through the Grafana webpage.
26:59
You can go sign up for a free
27:01
cloud account. You can download it for free,
27:03
and there's a bunch of contact info and
27:06
community info up there on that page. So
27:08
definitely welcome anybody to reach out and say
27:10
hello. Excellent, excellent stuff.
27:13
Well, listen, Aaron, you want to wrap it up, take us
27:15
home? Yeah, absolutely. So Ronald, first of all,
27:17
thank you very much for your time today. And on behalf
27:19
of myself and a very
27:21
time confused Brian, thank you very
27:24
much for listening this week. And if
27:26
you are out there, please tell a friend about
27:28
the show and please leave us a review wherever
27:30
you get your podcasts. And for that, I will
27:32
wrap for this week and we will talk to
27:34
everyone next week. Thank you
27:36
for listening to The Cloudcast. Please
27:39
visit thecloudcast.net to
27:41
find more shows, show notes, videos,
27:43
and everything social media.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More