Add Database Monitoring (DBM) query samples support

[veeshpath]

---

Feature Request

Add support for querying Datadog Database Monitoring (DBM) query samples via pup.

Context

pup covers aggregated DBM metrics (mysql.queries.*) well, but the per-sample data — including trace tags parsed from SQL comments — is only accessible through the DD web UI.

What's Missing

DBM query samples contain metadata not available in aggregated metrics:

• @trace.tags.* — application-level context (job name, source file, line number) extracted from SQL comments
• Full (non-normalized) SQL text
• Per-sample execution plan and timing

Proposed Commands

# Search query samples
pup dbm samples search --query "database_instance:my-db @trace.tags.job:*" --from 1h --limit 10

# Aggregate samples
pup dbm samples aggregate --query "database_instance:my-db" --group-by "@trace.tags.job" --compute "avg:@duration"

Use Case

Building CLI tooling for query performance investigation. The aggregated metrics tell us which queries are slow, but the trace tags tell us where in the application they come from. Currently requires switching to the DD web UI to bridge that gap.

[dujuku]

@veeshpath we’re working on adding DBM support to pup. In the meantime, have a look at this upcoming guide for how to retrieve query samples through the API: DataDog/documentation#35162

Can you provide a 3-5 sentence summary of what you’re building and why you want to use the data?

[veeshpath]

Thanks for following up so quickly!

In our application, we tag all queries with the line of code the generated it. I've so far built a tool and skill on top of pup for analysing the slowest queries on our database, but it's missing the application context to make it immediately actionable.

I'm happy to implement btw, if that's on the table