It would be extremely helpful if mcp-grafana could query an InfluxDB datasource—whether it’s a self-hosted node on EC2 or an InfluxDB Cloud bucket. Right now we have to write extra glue code just to perform a simple aggregation like min, max, and mean temperature across multiple sensors over the last 30 days.

Proposed approach

• New tool: query_influxdb
  • Inputs
    • datasource_uid – Grafana UID of the InfluxDB datasource
    • query – raw query text
    • start, end – RFC 3339 or epoch timestamps; omit both for an instant query
    • dialect – one of sql (default), flux, or influxql
  • Behaviour
    1. Build a payload for Grafana’s /api/ds/query, mapping query into rawSql, queryText, or the InfluxQL field according to dialect.
    2. Forward the request unchanged.
    3. Return the response JSON exactly as received (same pattern as query_prometheus and query_loki_logs).

Motivation

• Removes boilerplate for both self-hosted and cloud InfluxDB users.
• Keeps the tool signature consistent with existing datasource helpers.
• Supports all three InfluxDB query languages without breaking older setups.

Please let me know if this aligns with the roadmap. I'll be happy to contribute.