The 5 Levels of Self-Driving Production — Eric Schwartz, Traversal

AI Engineer · 18 min · 175 sentences · from YouTube's caption track

Each timecode opens YouTube at the start of that sentence. Line anchors (#s42) are the cue ids in the WebVTT, and every line carries its start and end seconds. All transcripts has every talk, and the whole corpus as one file.

00:01[music]
00:12Hey everybody.
00:13Um, my name is Eric.
00:15I'm a product manager at Traversal and today we're going to talk about self-driving production.
00:21So, just a quick primer on what I'll cover.
00:24Uh I'm going to talk a little bit about how AI agents are changing the software development life cycle.
00:30Uh the problems that that introduces for engineers and how AI for site reliability engineering fits in uh which is where traversal comes into play.
00:40And then I'll talk about a few examples of um what this looks like in practice for enterprises that we work with at Traversal uh and a bit about how we've built our AIS surre
00:53uh to deliver the results that we're seeing with Fortune 500 companies.
00:58So uh I guess to start some quick context on uh kind of the different phases of software engineering.
01:06Uh so sure you're all familiar but at a high level you could think about software engineering as kind of three buckets.
01:12There's system design like what is it that I want to build?
01:15What's the architecture that I want to build?
01:17There's development which is actually writing the code building the thing.
01:21And then there's troubleshooting.
01:22So once you build something and it's interacting with the real world, it starts to break uh and you need to fix it.
01:29And so with the advent of coding agents, the development portion has shrunk.
01:34development is a lot faster.
01:36Folks like me without any formal computer science training can now contribute u production code.
01:42And so this is great.
01:43Development's a lot faster.
01:45Teams can do 10x more.
01:47But then what happens to either end of the spectrum?
01:50What happens to system design and what happens to troubleshooting?
01:54The hope is that you spend more time on the design part thinking creatively about what it is that you want to build.
02:00But what we actually see with the enterprises that we work with is that um more and more time is being spent on troubleshooting.
02:08Uh much more code is being written.
02:11People have less understanding of the code that's being pushed into production.
02:14And so what you end up with is more issues, more comp uh more complex environments.
02:20Uh, and instead of spending all your time on the creative architecture and design, your teams are bogged down on troubleshooting.
02:29And the the data kind of bears this out.
02:32This is a big problem.
02:33Um, by some accounts, enterprises are spending upwards of $400 billion a year.
02:38Uh, 40% of executives say that this is a problem that their teams face.
02:43And on average, engineers are losing seven plus hours each week uh just troubleshooting when they're on call.
02:51And this is going to become an increasing issue that we'll all hear more about, that we'll all experience more over time as tools like cloud code and codeex and cursor,
03:00which are fantastic, um uh they are fantastic, but they result in more code and more complexity.
03:08And so what are we supposed to do about it?
03:10Um there are all kinds of great observability tools.
03:14Tools like data dog, elastic, splunk, service now where I worked for several years before joining traversal.
03:21Um these are great but uh there's only so much that they can do like these tools will tell you that something is broken.
03:28They'll point out all the things that have broken.
03:30Um they'll point out things that are maybe correlated but they won't tell you the root cause of the issue.
03:35And when I worked on observability products at Service Now, this was something I heard time and again from the customers that I served, which was like, you're telling me what's broken, you're not telling me why, and you're not telling me what to do about it.
03:47And so, no matter how many dashboards you create in these tools, it's not really helping teams keep up with the pace of innovation uh and complexity that's being added to their environments.
03:59And this is not just something we believe at Traversal.
04:02This is kind of a widely held view.
04:04um Google who kind of wrote the S sur handbook like the bible of site reliability engineering uh has commented on this directly and we also heard recently from anthropic
04:16u quotes where they described in their own words that LLM are not great at this problem.
04:21They're not great at pointing out what is the root cause of an issue.
04:27Now like why is this so hard?
04:29What makes this so challenging?
04:31Um in this particular example that we have on the screen like which was based on something we saw with one of our customers.
04:38Um if a checkout API is failing uh you might need to make five to 10 hops across like dozens of services and pabytes of data to actually get to the root cause.
04:50And so in a very small contained environment, perhaps there's a single seasoned SR at your company that knows the full stack and can debug everything and has all the tribal knowledge to sift through all your data.
05:02But if you're a Fortune50 enterprise or a Fortune 100 enterprise, there's no single engineer that has all of that context.
05:09And so what you end up with is a war room with 50 engineers like pulled in uh and many many hours trying to debug this issue because everyone has
05:19their own particular slice of context.
05:22No one has kind of the full scope and breadth of this.
05:25And so no, it's really challenging for a human or a human paired with cloud code to get kind of the full spectrum of context needed to jump from a checkout API is failing over here
05:38five hops away to the expired TLS certificate that maybe caused the issue.
05:46And so that's kind of what motivated u our founders to to launch traversal.
05:52Fundamentally the belief that we have at traversal is that uh root cause analysis is not an observability problem.
06:00It's a it's a causal problem.
06:02And so that's why our four founders, three of them come from academia.
06:06One comes from quant finance.
06:08Uh and they've dedicated like decades of their lives towards causal machine learning which is how do you find cause and effect?
06:18uh and have applied it to this problem because it's a massive one that enterprises are really struggling with.
06:25And so that kind of brings us to the topic of this talk which is self-driving production.
06:30Um what we're building at Traversal is the ability for an enterprise to basically have a self-driving production environment.
06:39What that means is a closed loop system where uh as issues break the AI system finds the issue the finds the issue or the incident or the alert finds the root cause of what caused that issue in the first place puts up a fix verifies that it's been fixed and the loop is closed with hopefully without ever
07:00having to page or disrupt any of your team members so that they can keep focusing on building.
07:05And this is a spectrum.
07:07U a lot of companies are not fully there yet.
07:10We like to think of it in terms of levels almost like self-driving cars uh or autonomous driving.
07:16We think about it in the same frame.
07:18So level zero is where most companies are today where everything is fully manual.
07:23You're pulling a bunch of people into a Slack war room or onto a Zoom call.
07:27You're spending hours debugging things.
07:30Level one might be where you have some rules and automations.
07:34And we're hearing a lot about like loops that teams are building now with tools like cloud code or cursor or codeex.
07:40Uh and so these might be like rulebased automations where you can build out a very structured specific workflow in response to an alert.
07:49Those rule-based automations kind of break down when it's a novel situation.
07:53And usually a sub one is something you haven't seen before.
07:56There's not really a runbook for it.
07:57And so the rule-based automation breaks down there.
08:01And level two and level three is kind of where you start to get a little more automation built in.
08:06In level three, perhaps you have like a a homegrown agent that is really good at debugging a issues for a specific service or a specific set of alerts.
08:17But really where we see the leap is going from level three to level four, which is cutting across an entire environment.
08:24So hundreds of services like hundreds of repos, thousands of lo log indexes.
08:30It's quite challenging to homegrow a solution that can handle that.
08:34And level five is kind of the holy grail, which is not only can we diagnose issues across a full production environment for a large enterprise, but also put up the fixes and verify the fixes as well.
08:47And so, uh, we're really proud that at Traversal, we get to work with and serve some of the biggest companies in the world.
08:55Companies like American Express, who I spend a lot of time with personally, uh, Pepsi, Digital Ocean, Capital One.
09:03Um, and it's really cool to see the volume and scale of data that Traversal is able to handle.
09:10So, you know, trillions of logs that are generated, trillions of spans, tens of billions of metrics and events, um, and hitting 80% plus root cause uh, for these high severity incidents at this scale is a technical feat that we're pretty proud of.
09:26And so, beyond just these kind of results at a high level, um, as a product manager, I like to think about the use cases that we build for our customers.
09:36And so there's there's a few different types of scenarios and use cases that we found that can deliver a lot of value.
09:42I'll dig into a few in more detail, but at a high level, there's alert intelligence, a product that I spend a lot of time on, which is helping on call teams
09:52triage and prioritize hundreds or thousands of alerts that they may be receiving in a single day, helping you deal with alert fatigue.
09:59There is incident root cause analysis, which is where we got started.
10:03How do we help you find the root cause, pinpoint the root cause of an incident in minutes when it otherwise would have taken hours and dozens of engineers?
10:11There is self-healing, which is not only telling you what the root cause was, but putting up a fix to address and mitigate that issue.
10:19There's production support, which is the the variety of questions and issues that your team might be dealing with in any given day, wanting to chat with their production environment,
10:30get specific data points that maybe aren't part of a a broader incident, but are still taking up a lot of time and toil for your teams.
10:38And then pre-production, there is code resilience.
10:40Like when you're putting up changes, how do you make sure that this code is making your system more resilient and reliable over time?
10:47And so at Traversal, we cover kind of the whole gamut of use cases.
10:51Um, but I'll dig into alert intelligence and incident root cause just to give you a bit of a more detailed sense of what that might look like, what that looks like, and I'm happy to answer any questions after about the others.
11:06So for alert intelligence, the example that I have here is uh based on our engagement with Pepsi.
11:13And so Pepsi uses traversal across a bunch of their different applications, but a very core one is for their supply chain.
11:21They have a lot of internal tooling and software to make sure that um the raw materials and the finished product can move through their supply chain.
11:30And a very critical piece of that is making sure that finished product gets from their warehouse to their trucks to their retailers at the end.
11:39And um before traversal their team would get thousands or tens of thousands of alerts every single week.
11:47Uh and at any given time a single engineer might have a backlog of 700 alerts.
11:52This is kind of a crippling state to be in.
11:55Like you have so much noise that you don't know where to look.
11:59You don't know what's broken and things slip through the cracks.
12:02Uh or you just try to keep up and burn out.
12:04And so this was a big problem that Pepsi was facing and they used traversal uh to help parse the signal from the noise of these tens of thousands of alerts.
12:15And so now instead of each engineer having a backlog of 700 alerts, they get a very pri a prioritized and filtered set of alerts that have already been pre-investigated.
12:27We flag to them the alerts that are actually worth digging into.
12:31We flagged to them opportunities to reduce alert noise, whether that's updating the underlying alert rules in their code, uh dismissing alerts that don't need to be addressed right now,
12:41uh or creating tickets for alerts that are maybe technical debt that can be handled later on but aren't super important to deal with at this moment.
12:50And so we've seen a lot of great results with Pepsi uh on this alert use case and many of our other customers as well.
12:57Um the next example is around incident root cause analysis and and this is a case study from our engagement with annex um where I've personally spent a lot of time myself.
13:08Um, so before working with traversal, what an incident looked like at American Express is anywhere from five to 10 teams get paged, anywhere from 20 to 50 engineers, and it could take, in this particular example, 60 minutes, but it could take hours or days
13:27to get resolution to to an incident.
13:30You can imagine if customers aren't able to pay their credit card bill, like that's a big problem.
13:34Or if they're not able to log into their mobile app, that's a big problem.
13:38And so now uh traversal is the first responder to every single incident that's created at American Express.
13:46And so what that looks like is the second the bridge is declared.
13:50That's the term they use internally.
13:51The second the incident is declared traversal is dispatched within 3 minutes we'll post a very detailed root cause analysis in the Slack channel where that incident is being managed.
14:02We can post updates to Service Now tickets if they want to as well.
14:06And rather than paging five teams and 58 engineers because you have that initial analysis, it's a lot less chaotic, you have a sense of what actually broke.
14:15And either no teams get paged or maybe you page the one or two teams just to verify the findings that traversal posted.
14:23Uh and so we save in this case like 50 plus engineers uh from the time and headache of being paged uh into this incident.
14:32Um, and that's especially appreciated when these incidents are happening in the middle of the night.
14:36Like this is a a good night of sleep that 53 engineers can have now.
14:42And so those are those are two of the use cases that we're especially proud of at Traversal, but like I mentioned, there's a bunch of others.
14:48Maybe in the last couple minutes I'll just highlight like AI site reliability engineering is a very hot space.
14:55You've probably, if you're familiar, you've probably heard of a lot of companies in the space or claiming that they're solving these problems.
15:03Um, we think that there's a few really key questions you should be asking uh that are critical to getting this right.
15:10So, the first is can your AI surre see all of your production data?
15:16If you're only giving it a slice or if it has gaps in its understanding, it's going to be very hard to get to a very detailed root cause.
15:24The second is can it search through this data can be it's often pabytes of data hundreds of billions of logs um without blowing up costs or taking down your observability
15:35infrastructure.
15:37Um this is really not trivial to do at scale for a Fortune 100 enterprise.
15:44The third is can it map out the relationship relationships between all of these entities in your data.
15:50even if you're able to ingest and read through all this data, if you don't know what to do with it and if you're lost, it's kind of worthless.
15:57And then four, does it get better and smarter over time autonomously without having to dedicate engineering resources to maintaining a bunch of markdown files or having a bunch of forward deployed engineers taking up your your team's time to maintain the system knowledge?
16:14And finally, u can it make multiple hops?
16:17Can it find the non-trivial, non-obvious root cause that's far away from the initial symptom in a matter of minutes?
16:24What we've seen is like anything more than 5 minutes and you've kind of lost the plot.
16:29You've lost the patience of the on call team and people will fall back onto their own habits.
16:34So, it needs to be really good.
16:35It needs to be really fast.
16:39And I won't go into this in too much detail, but this is kind of how we think about these five questions at traversal.
16:47So at the bottom of the stack is how we think about kind of a analyzing all your data uh without increasing cost.
16:56So basically how do we ingest and map all of the data uh that we're connected to and integrate with all of that data.
17:03In the middle of the stack is how we make sense of that data.
17:06And so we call that our production world model.
17:09How do we map out all the relationships between all the dent the data that we're ingesting?
17:14And then finally, how do we surface that and package that up in either use cases or user experiences that are valuable to your team?
17:24And so there's a lot of components to getting this right u especially at Fortune 500 scale.
17:32And so um that's a bit about what we've been building at Traversal, what we're delivering at Traversal.
17:38The core pieces like I mentioned are the production world model.
17:41So how we map and make sense of all the data that we integrate with and then what we call uh what we call refer to as our causal search engine.
17:49Basically the the harness and mechanism that we give our agents to search through all of that data uh that we've mapped out.
17:57And so the production world model and the causal search engine are how we believe we are going to help Fortune 500 enterprises get to the level five of autonomy
18:08fully self-driving production.
18:12Cool.
18:13Thank you for your time.
18:30[music]