Keep your stack. Add the layer it's missing.
Prometheus, Grafana and Datadog tell you what broke. Sentinel turns their alerts into consolidated service health, alert routing that behaves, uptime history and a status page your customers can read. Five lines of YAML, no metric leaves your cluster.
alertmanager.yml
receivers:
- name: sentinel
webhook_configs:
- url: https://app.sentinel-app.io/api/v1/signals/alertmanager?service_label=app
http_config:
basic_auth:
username: YOUR_KEY_ID
password: YOUR_SECRETForward your alerts in three steps
No agent to install, no migration, nothing to rip out.
Create an API key
One key from the Sentinel admin. Alertmanager and Grafana authenticate with it over HTTP Basic, so it works with the prometheus-operator CRD too.
Point your receiver at Sentinel
Add a webhook receiver to alertmanager.yml or a contact point in Grafana. Datadog posts to the regular signal endpoint with a custom payload.
Watch it consolidate
Alert rules become components of a service, firing and resolved become down and up, and your Prometheus labels become Sentinel labels.
What you get on top
The parts an observability platform leaves you to build yourself.
Alert routing that behaves
Retries with backoff, cooldowns scoped per incident that reset the moment a service recovers, and a circuit breaker that pauses a failing destination instead of hammering it.
A status page your customers read
Every forwarded alert feeds the same public status page and uptime history — the outward-facing job Datadog and Grafana were never built for.
Maintenance windows that really mute
Schedule one-shot or recurring windows. Matching alerts stop paging and stop counting against your SLA, whichever stack raised them.
One health view across sources
Prometheus alerts, Kubernetes workloads, CI pipelines and cron jobs land in the same service list, with the same up, degraded and down model.