wachd

Wachd is an open-source, self-hosted on-call and alert intelligence platform. It tells you why an alert fired, not just that it fired. When an alert triggers, Wachd automatically collects the last 10 commits to the affected service, 30 minutes of error logs, and metric history around the alert time. It sanitises all PII, runs AI root cause analysis, and delivers a plain-English probable cause to the on-call engineer — via SMS, voice call, Slack, or Teams — before they've even opened their laptop. Key features:

• AI-powered root cause analysis (Ollama for air-gapped/self-hosted, Claude or OpenAI for cloud)
• Flexible on-call scheduling — rotations, time windows, overrides, escalation chains
• Works with Grafana, Datadog, Prometheus, Alertmanager webhooks out of the box
• Per-team self-service configuration — teams onboard themselves in under 5 minutes
• Breach intelligence — daily CVE scan against your actual dependency inventory, auto-PR with fix
• Full REST API — every operation available in the UI is also available via API
• Multi-tenancy built-in — teams are fully isolated, no cross-team data leakage
• Helm chart for Kubernetes (EKS, AKS, GKE, k3s, on-prem)

Replaces: OpsGenie, PagerDuty, VictorOps

Self-hosted means your alert data, logs, and on-call schedules never leave your infrastructure.