Wachd
Self-hosted, AI-powered root cause analysis for on-call alerts from Grafana, Datadog and Prometheus.
🔗 Visit Wachd
Description
Wachd is what an experienced on-call engineer would do at 3am if they had superhuman speed: the moment an alert fires from Grafana, Datadog or Prometheus, it automatically pulls the recent commits, error logs and metric history around that moment, strips out any sensitive data, and hands the on-call person a plain-English explanation of what probably broke and what to do about it — before they've even opened their laptop.
It's self-hosted on Kubernetes, Apache 2.0 licensed for the open-source core, and supports fully air-gapped deployments using local Ollama models so no data ever leaves the company's network — a real option for regulated environments where sending logs to a cloud AI vendor isn't allowed. Paid SMB and Enterprise tiers add hosted convenience and support on top of the free self-hosted base.
💬 Our review
The short version: Wachd targets a genuinely underserved niche — AI root-cause analysis you can run fully air-gapped — and that's the reason to pick it over a bigger name like OpsGenie or Datadog's own AI features.
At 36 GitHub stars it's an early-stage project without OpsGenie's decade of on-call scheduling maturity or Datadog's all-in-one observability suite, so teams already happy with their alerting tool's routing and escalation features won't gain much by switching just for on-call scheduling. Where Wachd earns its place is the compliance angle: if your organization can't send logs or commit history to an external AI API, an air-gapped Ollama-based RCA tool is a narrow but real requirement, and Wachd is one of the few open-source options built specifically for that case. For everyone else, it's a promising but young tool worth piloting rather than betting the whole on-call stack on immediately.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Open source self-hosted tier free; SMB ~$99/month, Enterprise custom
Pros
Fully air-gapped option (local Ollama)
Automatic root cause correlation
Free open-source core
Kubernetes-native
Cons
Very early-stage (36 stars)
Less mature on-call scheduling than OpsGenie
Small integration ecosystem