ethpandaops Query Guide

This skill is a router. The full, always-current usage guide is owned by the code — read it first and follow it:

• CLI: panda getting-started
• MCP: read the panda://getting-started resource

It is generated live from the running server — workflow, the Python clickhouse/prometheus/dora/storage API, the ClickHouse raw-vs-refined rules, partition-key filters, sessions, and storage — so it never goes stale. Use panda docs <module> for a module's full API and panda search examples "<topic>" for worked queries. Everything below is the durable context that guide does not carry.

Discover names — don't hardcode them

Datasource, cluster, and table names are owned by the proxy and change over time, so enumerate them from the live tooling rather than pasting a name from memory, an old chat, or a screenshot: