Alertmanager
The deduplication, grouping, and routing layer for Prometheus alerts. Sends to PagerDuty, Slack, email, webhooks.
| License | Apache-2.0 |
| Repository | https://github.com/prometheus/alertmanager |
| GitHub stars (approx) | 8500 |
| Language | Go |
| Self-host deployment | single binary, docker, kubernetes helm |
| Commercial hosted option | no / self-host only |
| Freshness status | Fresh |
| Freshness checked | 2026-05-17 |
| Categories | alerts |
Alertmanager is the alert-handling sibling of Prometheus. Prometheus evaluates rules and fires; Alertmanager dedupes, groups, silences, and routes to notification backends (PagerDuty, OpsGenie, Slack, email, webhooks, Matrix, MS Teams).
Pick it if: you run Prometheus — Alertmanager is the canonical answer for everything downstream of “an alert fired.” Stable, boring, ~8500 GitHub stars, CNCF-graduated by association with Prometheus.
Skip it if: you want full on-call scheduling, escalation policies, or incident timelines — Alertmanager is a routing layer, not an incident management platform. Pair it with Grafana OnCall, Opsgenie, or PagerDuty for that. Also skip if you’re not in the Prometheus ecosystem; OpenTelemetry-native alerting (e.g., via SigNoz) may fit better.
Deployment: single Go binary or container. Helm chart bundled in kube-prometheus-stack. Config is YAML with a route-tree that takes a few iterations to internalize.