prometheus

What This Skill Does

The Prometheus skill is a powerful monitoring integration designed for OpenClaw users to interact directly with their observability infrastructure. It acts as a bridge between your natural language queries and one or more Prometheus monitoring servers. Whether you are managing a single local instance or a complex multi-server environment, this skill allows you to retrieve real-time metrics, system health, and performance data using intuitive CLI commands or automated agent responses. It handles complex configurations, including HTTP Basic Authentication, and provides a centralized way to query aggregated data across your entire infrastructure stack.

Installation

To integrate this skill, navigate to your terminal and use the following command: clawhub install openclaw/skills/skills/akellacom/prometheus. Once installed, initialize your configuration by running node scripts/cli.js init within the skill directory. This setup wizard creates the necessary prometheus.json configuration file, allowing you to map your various server endpoints and credentials. You can define multiple instances and set a default, ensuring seamless switching between your production, staging, and development clusters.

Use Cases

This skill is perfect for infrastructure engineers and DevOps practitioners who need immediate visibility into system health without switching context. Common use cases include: checking disk space availability across a server fleet, monitoring CPU and memory saturation during incidents, inspecting network throughput, and verifying the 'up' status of critical services. It is particularly effective for teams using federation or multi-region deployments, as the --all flag allows for instant cross-cluster comparisons.

Example Prompts

1. "What is the current CPU usage across all my production servers?"
2. "Check if there is any disk space pressure on the staging cluster."
3. "Show me the network traffic statistics for the main API gateway instance."

Tips & Limitations

To get the best performance, always leverage the prometheus.json configuration file rather than relying on legacy environment variables. If you are querying multiple instances, ensure your Prometheus servers are reachable from the host running OpenClaw, as network latency or firewalls can impact response times. Note that this skill requires read access to your monitoring endpoints; ensure that your service accounts have the appropriate Prometheus API permissions. Use the --all flag sparingly for complex queries, as it aggregates results across every configured endpoint, which may result in verbose output.