Language
This documentation is in English. Spanish version coming soon — contact us if you need help in Spanish.
centinelAI Documentation¶
Your SRE that never sleeps.¶
centinelAI absorbs all the noise from your infrastructure with AI, filters 99% automatically and only alerts when something is critical. Without changing your stack.
How it works¶
-
Receives alerts
Webhooks from Kubernetes, Prometheus, Grafana and GitLab. Set up in under 10 minutes.
-
AI analyses it
4 agents powered by Claude AI evaluate each alert in context. Score from 0 to 100 with a natural language explanation.
-
Only interrupts when it matters
Only score >70 reaches your team — with probable cause, impact and recommended actions.
-
Automatic postmortem
When an incident is resolved, Claude Sonnet generates the full postmortem with timeline and root cause.
Available connectors¶
-
Kubernetes
Python agent in your cluster. Detects CrashLoops, OOMKills, ImagePullBackOff and more in real time.
-
Prometheus / Grafana
Webhook receiver compatible with Alertmanager and Grafana Alerting. Zero-config with your existing stack.
-
GitLab CI/CD
Detects pipeline failures and correlates deploys with infrastructure anomalies automatically.
-
Slack
Rich notifications with action buttons. Declare incidents and snooze alerts without leaving Slack.
Plans¶
| Starter | Team | Pro | |
|---|---|---|---|
| Price | €0/mo | €99/mo | Custom |
| Services | 5 | 25 | Unlimited |
| Claude AI scoring | ❌ | ✅ | ✅ |
| Slack + Email | ❌ | ✅ | ✅ |
| AI Postmortem | ❌ | ✅ | ✅ |
Immediate ROI
A 2h incident with 10 engineers = €4,000. centinelAI Team = €99/month. Preventing 1 incident per month → ROI of 3,940%.