Prometheus Connector
OnPremiseAgent exposes a Prometheus-compatible metrics endpoint for comprehensive monitoring of your AI agent fleet. Track query latency, token usage, error rates, compliance violations, and resource utilization. Combine with Grafana for dashboards and Alertmanager for intelligent alerting — all within your infrastructure.
TokenService Account
Auth
Token, Service Account
Category
Observability
Compatibility
Prometheus, Thanos, Cortex, Mimir, VictoriaMetrics
Tier
Available
Purpose-built capabilities
Everything you need to integrate Prometheus into your on-premise agent workflows.
Metrics Endpoint
Standard /metrics endpoint exposing 50+ metrics including query latency, token usage, and error rates.
Custom Metrics
Define custom metrics for business-specific KPIs like compliance scores and agent accuracy rates.
Alert Rules
Pre-built Prometheus alert rules for common scenarios: high latency, error spikes, token budget exceeded.
Service Discovery
Automatic service discovery for Kubernetes deployments — new agents are scraped without configuration changes.
01
Enable Metrics
Enable the Prometheus metrics endpoint in your OnPremiseAgent configuration (enabled by default on port 9090).
01
Key Benefits
Why enterprises choose this connector
- 50+ built-in metrics covering all aspects of agent operation
- Pre-built alert rules for immediate monitoring coverage
- Kubernetes service discovery for zero-config scraping
- Compatible with all Prometheus-compatible tools (Thanos, Cortex, Mimir)
- Custom metric support for business-specific KPIs
01
SLA Monitoring
Track agent response times and availability against SLA targets with automatic alerting on violations.
02
Cost Optimization
Monitor token usage per agent and department to optimize costs and enforce budget limits.
03
Compliance Monitoring
Track compliance violation rates and policy enforcement metrics for regulatory reporting.
Frequently Asked Questions
What metrics are exposed?
Query count, latency histograms, token usage, error rates, active agents, compliance scores, and resource utilization — plus custom metrics you define.
Does this work with Thanos or Cortex?
Yes. The metrics endpoint is fully compatible with any system that scrapes Prometheus-format metrics.
Works great with
Combine Prometheus with these connectors for a complete integration stack.
Ready to connect Prometheus?
Deploy on your own infrastructure with full data sovereignty. Get started in minutes.