NOT A DEMO — Be a Guest SRE in Our Real Environment
We replaced our runbooks with agents. Want to try them?
At SREDay London, we handed the keys to our test environment to a room full of strangers. Not a sandbox. Our actual developer environment — the same clusters, namespaces, and services that the RunWhen engineering team pushes to every day.
Nobody in that room had seen our system before. They walked in, sat down, and started operating.
It worked so well we’re turning it into a standing invitation. We figure if we can’t hire enough SREs, we’ll just crowdsource them.
Sign up for a Guest On-Call session.
What you get
- The same agent permissions our team has. Remediations are real. Tickets you create are real.
- A RunWhen engineer on the call. These are our real environments — we need to make sure you don’t take us down.
- A library of runbook tools built over three years, and agents tuned over about two weeks to wield them.
No cost. No contract. No pitch.
What you’ll be operating
Our developer test infrastructure. Multiple Kubernetes clusters, hundreds of monitored resources — databases, queues, search indexes, API gateways, certificate management, GitOps controllers. Over 600 health checks running continuously. When something drifts, the agents know before you do.
The agents don’t just observe — they act. Restart deployments, rollback versions, scale services, resize storage. When an alert fires, an agent triages it and runs the right remediation. You’ll have the same access.
There are live issues in this workspace right now. Certificate renewals not completing. Scaling limitations under load. Timeout errors propagating through services. The kinds of real, messy problems your team probably deals with too.
You’ll discover the details when you get in there. That’s the point.
What SREDay London looked like
We co-presented with Lydia Thomas, who had never seen the system. No script. No safety net:
“Absolutely NO gating or safety in this demo. I can think of so many times at 2am in war rooms, where I can’t remember what the internet even is, and I’m so tired that I need onboarding back onto my own product, where I could’ve only dreamed about having a tool like RunWhen.”
Why we’re doing this
Most vendor demos are theatre. You nod politely and leave with a vague sense of what the product might do. You learn nothing about whether it would work at 2 AM when everything is on fire.
We think the only honest way to show what we’ve built is to let you use it. Against problems we didn’t stage for you.
I’ve written before about the surprising diversity of context that AI agents actually use in production — service relationships, event patterns, operational notes that senior SREs jot down between incidents. This session is your chance to see whether agents with all of that context can help someone with zero prior knowledge get productive in under an hour.
If they can’t, that’s important for us to know. If they can — well, that’s important for you to know.
Sign up
Pick a time on our calendar. 45 minutes. Orientation, issue triage, then choose your own adventure. A RunWhen engineer will be on the call the whole time.
Bring your skepticism. We’ll bring the live infrastructure.
And if you find a bug, congratulations — you just did free QA for us. We told you we were crowdsourcing.