Wachd
wachd.io
- Free plan
- Web
- API
- Self-hosted
What Wachd is
Wachd is an open-source incident alerting and diagnosis service for DevOps and SRE teams. It receives monitoring alerts, gathers related operational context, and sends an AI-generated root-cause diagnosis to the on-call engineer. For an alert, it can collect the last 10 GitHub commits, 30 minutes of logs from Loki or Datadog, and metric history around the alert period. AI choices include in-cluster Ollama or vLLM, OpenAI-compatible models, Claude, OpenAI, and Gemini. Scheduling supports primary, secondary, and custom rotation layers, time restrictions, and self-service overrides. Alerts can go by SMS, voice, email, or Slack, with escalation when unacknowledged and cancellation after resolution. Wachd says it runs in the customer's Kubernetes cluster, where incident data can remain in the VPC; air-gapped use is possible with an in-cluster model. It says it removes emails, IP addresses, API keys, and account IDs before data reaches the AI backend. The self-hosted Apache 2.0 plan costs nothing, with no paid tier or license fees.
Who it suits
Wachd is aimed at DevOps and SRE teams that want alert context and AI-generated diagnosis routed to an on-call rotation. Its in-cluster deployment may fit teams seeking to keep incident data within their VPC.
Good to know, and what to watch
- Collects commits, logs, and metric history around alerts
- Supports several AI backends
- SMS, voice, email, and Slack notifications
- Offers layered on-call rotations and escalation
- Free, self-hosted, Apache 2.0 licensed
- Requires Kubernetes deployment
- No paid tier is offered
- AI data handling depends on selected backend
Plans
Every plan and priceFree & open source
Free plan
The verdict
Wachd brings alert context, diagnosis, rotation scheduling, and multiple notification channels into one self-hosted service. Teams should consider its Kubernetes deployment and chosen AI backend when assessing fit.
What the maker publishes
- Access controls
- It lists Entra ID, Okta, and Google Workspace group mappings, with local accounts available as break-glass access. source
- AI backends
- It supports in-cluster Ollama, vLLM, or OpenAI-compatible models, as well as Claude, OpenAI, and Gemini. source
- Context collection
- It collects the last 10 GitHub commits, 30 minutes of error logs from Loki or Datadog, and metric history around the alert window. source
- Data residency
- The site says Wachd runs in the customer's Kubernetes cluster and incident data can remain inside its VPC, including in air-gapped deployments using an in-cluster model. source
- Deployment
- The site describes a Helm chart for Kubernetes and says deployment takes under 30 minutes. source
- Infrastructure access
- Wachd describes its access to GitHub, GitLab, logs, and metrics as read-only or query-only. source
- Integrations
- The site lists Grafana, Datadog, Prometheus, Loki, PagerDuty webhooks, GitHub, GitLab, Slack, Microsoft Teams, Twilio, and several AI inference backends. source
- Intended users
- Wachd says it is built for DevOps and SRE teams. source
- License and pricing
- Wachd states that it is free and open source, has no paid tier or license fees, and is licensed under Apache 2.0. source
- Notifications
- Notifications can use SMS, voice calls, email, and Slack, with escalation for unacknowledged alerts and cancellation when resolved. source
- On-call scheduling
- Schedules support primary, secondary, and custom rotation layers, time-window restrictions, and self-service overrides. source
- PII protection
- Wachd says it strips emails, IP addresses, API keys, and account IDs before data reaches the AI backend. source
- Purpose
- Wachd receives monitoring alerts, gathers related context, and sends an AI-generated root cause diagnosis to the on-call engineer. source
- Support
- The site directs questions to [email protected] or its Discord community. source
Checked against the maker's pages on 8 Oct 2026.