Integration · Monitoring
Prometheus + OrchestrAI
Catalog exported 2026-09-02 · Prometheus website
Ask Prometheus questions in plain language: instant and range PromQL queries plus a view of firing alerts.
OrchestrAI exposes 3 Prometheus operations: 3 are low-risk (read-only or low-impact).
What teams use it for
Rather than remembering PromQL, an on-call engineer asks for the 95th percentile request latency of the api-gateway over the last hour, or the memory usage trend for the ingest pods since the deploy, and OrchestrAI writes the query, runs it against your Prometheus endpoint, and summarizes the numbers. Pairing a query with the current list of firing alerts gives a quick answer to whether an alert is still justified. All three operations are read-only, and there is nothing yet for silencing alerts, editing recording or alerting rules, or inspecting scrape targets, so those remain in Alertmanager and your rule files.
Every Prometheus operation, with its risk level
| Operation | What it does | Risk | Step-level approval |
|---|---|---|---|
Get Prometheus Alerts |
Get active Prometheus alerts | Low risk | No |
Prometheus Query |
Execute instant PromQL query | Low risk | No |
Prometheus Range Query |
Execute range PromQL query | Low risk | No |
Risk tiers come from the catalog: low is read-only or low-impact, medium creates resources and is reversible, high modifies existing resources, destructive may lose data. Every plan that creates or changes resources is shown with its cost estimate and waits for your confirmation. Operations marked with a step-level approval pause again on their own step. Destructive operations require a typed risk phrase.
What you connect
A Prometheus credential (stored as prometheus).
Connected-service tokens are envelope-encrypted with a per-record key wrapped by a cloud KMS.
Prompts that work
- What is the p95 latency for the api-gateway service over the last 30 minutes?
- Show me the CPU usage of the ingest-worker pods in the prod namespace for the past 6 hours
- Which Prometheus alerts are firing right now?
Before anything runs
Every mutation shows its plan, cost estimate, and blast radius, then waits for your confirmation. Destructive operations require a typed risk phrase. Credentials are minted per run through OIDC federation and discarded afterward; nothing you create here is invisible later, because every resource lands in the desired-state ledger where drift is detected and can be converged. Details on the security page.
Frequently asked questions
- Does OrchestrAI write the PromQL for me?
- Yes, You describe the metric and time window, and it composes the PromQL for prometheus_query or prometheus_query_range, shows the expression, and returns the result.
- Can OrchestrAI silence a Prometheus alert?
- No, prometheus_get_alerts reads active alerts but there is no operation for silences or rule changes. All Prometheus coverage is low risk and read-only.
- How does OrchestrAI authenticate to Prometheus?
- You add a Prometheus credential once in the connections screen. It is envelope-encrypted with a per-record key wrapped by a cloud KMS and is only decrypted inside the run that needs it.
Related integrations
Try it on your own account
Connect your cloud read-only and see your resources, drift, and costs before anything runs. $5 minimum to start. Unused credits refunded in your first 14 days.
Unused credits refunded in your first 14 days.