Description
Wachd is a self-hosted on-call tool that tells an engineer why an alert fired, not just that it fired. It receives alerts from Grafana, Datadog and Prometheus through a webhook, then collects context before the page reaches anyone: the last ten commits from GitHub, thirty minutes of error logs from Loki or Datadog, and metric history around the alert window. An AI step analyses that causal timeline and writes a plain-English root cause with a suggested action. The model can run inside your cluster through Ollama, vLLM or any OpenAI-compatible server, or be Claude, OpenAI or Gemini. Personal data is stripped before the model sees anything. Notifications go out by SMS, voice call, email and Slack in the order you configure, with escalation if nobody acknowledges, and they are cancelled once the problem is resolved. Schedules support primary, secondary and custom rotation layers, time windows for business hours, nights and weekends, and self-service overrides. Each person can set their own notification rules. Wachd runs on Kubernetes and can work air-gapped, with single sign-on or local users, multi-tenant isolation and CVE tracking. The code is licensed under Apache 2.0 on GitHub, and a live demo is available.…
Preview
Highlights
- AI root cause analysis for Grafana, Datadog and Prometheus alerts
- Automatic context from GitHub commits, logs and metrics
- Rotations, overrides and time-window schedules
- SMS, voice, email and Slack with escalation
- Self-hosted on Kubernetes and air-gapped capable
- Apache 2.0 open source
Ratings & Reviews
No ratings yet
Be the first to rate Wachd.
People also visit

monitoring-expert
Configures monitoring systems, implements structured logging pipelines

app-observability
Get RED metrics + service maps + frontend RUM + AI/LLM monitoring out of Grafana

agent-platform-alert-configuration
Configures best-practice alerting policies for AI agents using OpenTelemetry
Wachd FAQ
What is Wachd?
A self-hosted on-call tool that receives alerts from Grafana, Datadog and Prometheus and explains the likely cause of each one.
Does my data leave my cluster?
The site says your data never leaves your cluster. Personal data is stripped before AI analysis, and you can use a model that runs in your cluster, such as Ollama or vLLM.
Which AI models can it use?
Ollama, vLLM or any OpenAI-compatible model running in your cluster, or Claude, OpenAI or Gemini.
Is it free?
Yes. It is open source under Apache 2.0, and the site says you can deploy it for free on Kubernetes.
Can it run without internet access?
Yes. The site says it is fully air-gapped capable.
What are the best Wachd alternatives?
Popular alternatives to Wachd include Spectate, iLert and AiSOC. See all Wachd alternatives




