Using Metricly to answer analytical questions

You are connected to a Metricly MCP server. The data layer the user cares about lives there. The conventions below come from real production traces; departing from them produces misleading numbers.

1. The discovery pattern

For any analytical question, start with two calls:

1. list_metrics — every metric in the org, with descriptions.
2. list_dimensions — every groupable dimension, with descriptions.

Then call query_metrics(...). Don't skip the catalog; metric names are not self-describing, and "obvious" metrics like total_revenue differ across orgs in unit, source, and standing filter rules.