Key concepts¶
Alert score (0-100)¶
| Score | Status | Action |
|---|---|---|
| 0-29 | 🟢 UP | No notification |
| 30-69 | 🟡 DEGRADED | Dashboard only |
| 70-89 | 🟠 DOWN | Slack + Email notification |
| 90-100 | 🔴 CRITICAL | Urgent notification |
Alert groups¶
The deduplicator groups similar events from the same service in 5-minute windows. 50 events from the same CrashLoop generate only 1 notification.
Services¶
centinelAI automatically detects your infrastructure services
as soon as it receives the first alert:
- Kubernetes pod: production/api-payments
- Prometheus alert: HighMemoryUsage
- GitLab project: myorg/api-payments
Incidents¶
An incident is a manually declared event (from Slack or the dashboard) that groups all related alerts. When you resolve an incident, centinelAI generates an automatic AI postmortem.
Plan and AI access¶
| Feature | Starter | Team | Pro |
|---|---|---|---|
| Price | €0/month | €99/month | Custom |
| Services | 5 | 25 | Unlimited |
| Rule-based scoring | ✅ | ✅ | ✅ |
| Claude AI scoring | ❌ | ✅ | ✅ |
| AI correlator | ❌ | ✅ | ✅ |
| Slack notifications | ❌ | ✅ | ✅ |
| AI postmortem | ❌ | ✅ | ✅ |