What it does: Connects to monitoring systems (Prometheus, Alertmanager, as a first step)
When an alert fires, it can either: - start an auto-investigation - wait for a human engineer to trigger the investigation Summarizes findings in Slack or UI (like “CPU throttling on service X after deploy Y”). Suggest remediation steps
Why we think this is interesting: Current AIOps tools feel heavy and enterprise. We wanted something that feels like adding another engineer to your Slack channel.
It focuses on MTTR reduction, reducing on-call fatigue and engineers' toil, rather than just more alerts or dashboards.
It’s built on the new wave of agentic AI frameworks, which makes multi-step investigations possible.
Right now, it’s in early beta. You can try it with: demo [https://www.opsworker.ai/en/learning-center/pre-mvp-showcase...] Expecting a major release next week
We’d love feedback from the HN community:
- Would you trust an AI teammate in your incident workflow? - Which integrations (Kubernetes, AWS, Grafana, etc.) would be most valuable to you? - Have you tried something similar? What worked, what didn’t?
We’re around to answer questions and share technical details (architecture, AI stack, integrations).
Thanks for taking a look!