Full Stack Observability Unites NetOps and DevOps
Full stack observability promises a complete view of the technology stack. In practice, network issues and application issues still live in separate silos. Teams waste time arguing over whether the app or the network caused an outage.
Sigi’s Brad Collins and Anthony Stefanelli examine that challenge with Alan Shimel at KubeCon + CloudNativeCon in Salt Lake City. Collins is a sales engineer, and Stefanelli works on go-to-market.
Correlating Network, App and Infrastructure
Sigi comes from the team behind ElastiFlow, known for deep network telemetry. The new platform is AI-native and built on OpenTelemetry.
It correlates network, application and infrastructure signals, including the Kubernetes CNI layer. That overlay gives teams the context behind any internal or external service.
Stefanelli says the stakes are rising. AI helps teams ship code faster, which means more issues reach production. NetOps and DevOps need to resolve them together.
Moving From Data Rich to Action
Collins has worked in monitoring since before anyone called it observability. Collecting all the data is now the low bar. Storing it has become expensive, and hardware prices keep climbing.
His answer is a shift from data rich to action. When the platform spots a condition of interest, it keeps the relevant data while teams troubleshoot. Afterward, it summarizes the issue and discards the rest.
The lessons learned are stored long term. That institutional knowledge helps whoever handles the next incident, so teams stop solving the same problem twice.
Conversational Observability
Adaptive sampling and an upcoming real-time rewind feature make relevant data easier to find. Conversational observability then brings the right people into a shared triage workspace.
Network, platform, SRE and systems engineers can work the same problem with full context. With full stack observability in place, troubleshooting becomes a team sport rather than a handoff between silos.
Sigi launched its beta this week as a cloud-based SaaS platform. General availability is planned for January, and the team hopes to offer an on-premises option.
Transcript
Hey everyone, we're back here continuing our day three coverage of KubeCon. It's been a great three days. So much learning, so much innovation, a lot of AI agents.
Let me introduce you to our next two guests here. I've got to my far left, Brad Collins. Yep.
And to my immediate left, the man in the middle, Anthony Signorelli? Stefanelli. Stefanelli.
Close. Don't ever talk to us again. Yeah.
Stefanelli. Yeah, it's not that hard. Brad, Anthony, welcome.
Thank you. Thank you. Brad, I'm going to ask you, maybe you could introduce yourself to the audience, give them a little bit about yourself.
Absolutely. Again, my name's Brad Collins. I'm a sales engineer here at Siggi.
Come by a long way of observability companies. Been in the industry for a few gray hairs. From before they called it observability.
Way before, yeah, back when we were holistic monitoring or a monitoring framework, if you will. Uh-huh. Right?
Yeah. Yeah. So I've been around for a few minutes.
Have a long history with our company founder, back to IBM Micro Focus days- Really? I remember ... if that sound makes sense maybe for some of the audience.
So yeah. Anthony, how about you? Yeah.
Thanks for having us, by the way. No problem. This is awesome.
The event's been great. But yeah, Anthony Stefanelli. Also, I'm on the go-to-market team for Siggi.
Really my background was based in, started more on network monitoring side in terms of my tech background, and then ventured off into the dark or bright side, depending on who you ask, of DevOps. And so, yeah, really excited to be here. I've spent most of my tech selling career in the startup ecosystem, so intimately familiar with early stage software companies that are taking awesome new ideas and products to market.
Very cool. So let's, I'm going to assume a lot of people in our audience don't know or are not familiar with Siggi. Sure.
Right. Give us the Siggi story. Yeah.
I'd say a little bit of a background for us, Siggi, we're a, not spinoff, but our parent company is a company called Elastiflow, which is a widely known- Our audience knows Elastiflow ... yeah, network observability, performance monitoring solution. And so our specialty there is our deep telemetry and depth in that area.
And so for us, one of the biggest challenge we've come to understand over the last several years with our customers is there's issues with the network and then there's issues on the application or infrastructure side. Right. It's a very siloed thing where the communication isn't great.
There's lots of gray area in terms of, well, what caused these issues? Was it the app or is it the network? Then now we got to pull in systems people.
There's not exactly a deep expertise across multiple domains. And so Siggi was born to solve three main challenges. The first one being, you should have a full understanding of your tech stack, full stack observability, which a lot of folks out here are saying full stack observability.
Not to hit you with another buzzword, but our version of that is, look, you should have a full understanding and your telemetry for your CNI network application and your infrastructure. And so what we do, we are a AI native, OpenTelemetry native, observability platform that will correlate your network application and infrastructure telemetry, overlay that so you can fully understand and have the context behind any given service, whether it be internal, external applications, to really have a holistic view. Okay, how do these three each domains play a part in getting this end-to-end desired result in terms of your infrastructure and technology?
And so, yeah, it's been really awesome to talk to the folks here at KubeCon. Had a lot of really good conversations. I think one thing that stuck out to us quite a bit is the DevOps community, obviously the CNI network's important, but sometimes it's hard to grasp how important it actually is.
Especially as your speed to launch and shipping things really fast is a huge thing nowadays, especially with the explosion of AI. And with that comes the burden of, well, there's probably more issues. People are shipping code faster.
There's more issues with applications. And so, okay, how do we combat that? How do we help engineers out with meantime to identifying issues, resolving them?
And so yeah, that's the birth of Siggi, is to solve that, hey, NetOps and DevOps should work more cohesively together. We should give them a platform where both of them can understand their actual problems and collaborate and solve them together quicker. Great.
Brad, we got to hear from Brad a little bit. So Brad, excuse me, my voice is going. I only got three more interviews for the day- It's all right ...
so hopefully I'll make it through. You can do it. But you mentioned you've been in observability a long time, so it's before we called it observability.
It's funny, you've been around the block. No matter how big the hard drive was, it seemed our software bloated up to it. It did.
It's the same thing with observability. As our capability to have a wider lens to take in more data for observability has increased, well, so has the flow of stuff we have to observe. AI agents being the latest, greatest part of this, where, man, we just can't get a break in terms of how do we scale?
And that's a big part of what you guys are about. Yeah. So we've moved away from the core capability being, can I collect all the data?
Yeah. That's the low bar of entry now. Yeah.
Can you get all the data in? Yes. Now we have challenges of what's it cost to store all that data?
So we're moving from being data rich to action. So we want to really change it to a paradigm of action where we can come in and say, we've seen an issue, a condition of interest, an insight, or however you want to call it, and come in and say, "Hey, you know what? " Then most importantly though, what we learn from that, we want to keep that institutional knowledge, that operational knowledge, and store it long term to make it accessible by who may be here next troubleshooting issue- Exactly ...
right? Yep. So yeah, we're trying to combat high storage prices.
We're trying to combat high hardware prices. Hardware these days, go buy some memory if you can afford to take a second lien out on your house, right? Sure.
So, yeah, those type of things. And even then you may not be able to buy it. Not very much.
There's memory sold till 2028. That's right. It's crazy.
Yeah. Anthony, you mentioned DevOps and all that. com.
We're cloud native now. com. The lines have blurred.
Today you can't do cloud native without DevOps, but you can't do DevOps really without cloud native. Right. And GitOps and how that fits in, and all of this runs on these platforms now that allow DevOps engineers and everyone to go faster.
We've seen this, I call it a grand unification theory kind of thing, where all these, SRE another one, security, all of these disciplines that used to be silos. And in some ways that's really what DevOps was about. Bringing all these silos down.
All of these are moving in now into the platform. Observability is part of it. So in some ways it's almost like, be careful what you wish for.
This is what you wanted, isn't it? And now we just got it, to your point, Brad, the low bar is just sucking all that information in, storing it, turning it into actionable intelligence, making it productive, profitable. Is the key to game, right?
And that's what you guys are about. Yeah. And I think with that evolution of, hey, everybody says, "Let's break down the silos.
" Okay, well, how does that actually play out in practice, right? Right. Because the first thing that typically happens when people go to try to achieve this collaboration across different domains and teams is, well, all of our data is in different areas.
Yeah. So when things occur and we need to all, "Hey, what's the context here? How do we resolve this?
" We have a pipeline crawl issue where I need to go find the data first, and that's the number one challenge that you're going to face. And then by the time you've actually collected all that data and found everything that you needed, okay, well, how do I piece this together? I need very specific expertise in each one of these domains, whether it's network or platform or DevOps or SRE.
And so a big thing for us when Cigi was being born was, okay, number one thing, we have this big issue, let's make it easy to find the data. And so to Brad's point, whether that's our adaptive sampling that we do or something called real-time rewind that we're going to be rolling out, it's like, okay, hey, let's easily access the data that matters when you need it, whether it's in your domain or not. And then we're going to have this big push for conversational observability.
And what we mean by that is we have essentially, as our platform can produce these insights for you, conditions of interest, as Brad said, okay, we need to pull in Mike from NetOps, Jerry from DevOps, Suzy from the SRE team, and maybe Rick from Systems because they've all had some touch or feel to whatever. So it's like your checkout service or whatever it may be, right? Let's pull all those people in, let's give them a triage workspace, right?
A triage room where they can collaborate and solve these issues together, given all the context from each of those data pipelines where we had to gather that information across network, app, or infrastructure. So I think to your point, though, when you said, "Hey, this is what you wish for," I immediately thought, yeah, everybody, it sounds great to work together, but how do we get all these tools to talk to each other? That was why- And that's why, yeah, OpenTelemetry is extremely huge for that.
Well, OpenTelemetry really helps. Yeah, and get to that. And tools like Hotel.
Yeah. Let's focus in on Cigi now, though. Pivot a little bit.
" What's the on-ramp? How do they do that? Yeah.
io, C-I-G-I dot io. We launched the beta this week. Mm-hmm.
We'll be launching our, GA should be early in the new year, in January. So yeah, please go check us out. We have the beta.
So GA is January, but beta's available now. Got you. Beta's available now.
Yeah. They could download it, get it up and running, write you guys for any help. Yeah.
It's a cloud-based SaaS platform. As of now, hopefully, we'll have plans to offer our on-prem solution. A lot of our experience in the past is with larger, more complex organizations and that might have sensitive data regulations, right?
So, a big thing that we've heard from a lot of the folks around here at KubeCon is how important, hey, these observability tools are really awesome, but my company needs on-prem. Yeah. That's not really an option for us.
Well, no, but you know what? So that is one of the... I don't want to call it a benefit, but one of the things you're seeing as a result of the AI is a lot of companies are saying, "I don't want to keep my stuff up in someone else's model maker or someone's cloud.
I want it local. " And for a long time, we were using the service closet to hang coats, but we'll start putting computers back in there. I do think that is a trend we're seeing.
Yeah. I'm not saying the cloud's going away or anything like that. Yeah, no.
You need to offer both. Right. You need a hybrid world.
Anyway, hey guys, I appreciate you coming on here. Awesome. What was the domain again?
Cigi, C-I-G-I dot... io. io.
S-I-G-I. Got you. Yeah.
io. Hey, new company launching. You watched it here first on Texteron TV.