AlertiGate
alertigate.com- Category
- Developer Tools
- Rank
- No. 2140Tools index
Previous survey · No. 2067 ·
- Type
- TOOL
- Use case
- Deployment & Operations
- Date
About
AlertiGate is a self-hosted tool that investigates every Kubernetes alert from Alertmanager or Grafana by reading metrics, logs, cluster state, GitOps deploy history, and a team's own runbooks, then posts an incident report to Slack with a root-cause finding, confidence score, evidence timeline, and a reviewable fix diff it never applies automatically. It runs as a single read-only container with no write permissions beyond get and list, and reuses past investigations of the same alert.
What it does
The site says AlertiGate takes an alert by webhook, has five agents gather evidence in parallel from Prometheus, Loki, the cluster, Argo CD and a wiki over MCP, then has one agent weigh hypotheses. It says the resulting report is rendered without a model call and sent to Slack, Teams or a signed webhook.
Stated on the product site
- Install
- Helm chart published on GHCR
- Cluster permissions
- ClusterRole limited to get and list
- Model providers
- Any OpenAI-compatible endpoint, with OpenRouter when unset
- Storage
- One SQLite file on a persistent volume
- Scaling
- Single replica only in this release
- Not yet built
- Jira and GitHub context, Slack replies, automated remediation, alert grouping, multi-cluster, Flux
Not stated on the site
- The page does not state a price or a software licence.
- Token cost is given for deepseek-chat on the vendor's lab cluster only; the page gives no figures for other models.
- The page does not show how accurate its root-cause findings have been across real incidents.
Written from the product site at alertigate.com.
Tags
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.