Integration · Metrics and dashboards
Scrape Opservo metrics with Prometheus
Scrape every Opservo server metric into Prometheus from one URL per workspace or per server, with an API token. Gauges named opservo_*, labelled by host.
In short
Every metric Opservo collects is available in the Prometheus text format from one URL per workspace, or one per server. Point a scrape job at it with a workspace API token and your existing Prometheus, Grafana and alert rules keep working. Metrics are gauges named opservo_<key>, labelled with the agent id, hostname and display name, and change at most once a minute.
What you get
Your Prometheus stack gets the numbers; Opservo keeps the explanations.
- Fleet endpoint: GET https://api.getopservo.com/api/v1/metrics/prometheus.
- Per-server endpoint: GET /api/v1/agents/<agent-id>/metrics/prometheus.
- CPU, load per core, memory, disk per mount, service up/down, container memory, MySQL connection use, SSL days left, health score, open critical alerts and agent online state, among 80+ metrics.
- Per-item metrics carry one extra label such as mount, service, container, instance or domain.
How to set it up
Create an API token with the metrics scope and add one scrape job.
- In Opservo, open Account → Developer → API tokens and create a token. It starts with opv_ and is shown once.
- Add a scrape job with scheme https, metrics_path /api/v1/metrics/prometheus, target api.getopservo.com and the token as the authorization credential.
- Scrape every 60 seconds; the agent reports once a minute, so faster scrapes return the same values.
Good to know
- Per-mount, per-service, per-container, per-instance and per-domain values are latest only; CPU, memory and load also have history through the query endpoint.
- Prefer push? A Metrics stream channel sends the same metrics to your endpoint, Datadog or New Relic.
- Revoking the token stops scrapes at once with HTTP 401.
Prometheus and Opservo, answered
Does Opservo export Prometheus metrics?+
Yes. A workspace API token unlocks a Prometheus text-format endpoint for the whole workspace or for one server, with 80+ gauges named opservo_*.
How often do Opservo Prometheus metrics change?+
At most once a minute. The agent sends a heartbeat every 60 seconds, so a 60-second scrape interval is enough.
Read next
Free for 2 servers, no card. Install the agent in about 30 seconds, then connect Prometheus.