feat(scm-multi-platform-detection): Adding new metrics

[Abdkhan14]

Motive:

• Seems to be redundant to be fetching the entire hierarchy of entries. We can use simple graphql queries with capped depth to get entries TILL that depth
• Identify areas that could use some optimization

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 6c23be0. Configure here.}

[cursor]

Unbounded per-path metric emissions

Medium Severity

Each detection run loops over every path in needed_paths and calls sentry_sdk.metrics.distribution once per path. That set is uncapped and can include every matching manifest or match_ext file in a large tree, so a single request may emit thousands of metric samples and add noticeable latency on top of an already expensive GitHub tree fetch.

 

^{Reviewed by Cursor Bugbot for commit 6c23be0. Configure here.}

[Abdkhan14]

This is okay. Max needed length over the last 2 days was 41, with a p90 of 4

[sentry]

Bug: The loop over needed_paths emits a separate metric for each path, which can cause high metric volume and unnecessary overhead in large repositories.
_{Severity: MEDIUM}

Suggested Fix

Instead of emitting a metric for every path in needed_paths, aggregate the data first. For example, you could calculate the maximum or average path depth and emit a single metric. This would capture the intended insight without the performance overhead of numerous individual calls.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: src/sentry/integrations/github/multi_platform_detection.py#L641-L645

Potential issue: The code iterates over `needed_paths`, an uncapped list of file paths,
to emit a metric for each path's depth. In large monorepos with many packages (e.g.,
50+), this can result in a high volume of `sentry_sdk.metrics.distribution()` calls for
a single platform detection run. While each metric call is individually lightweight and
non-blocking, the aggregate overhead from dozens of calls is unnecessary and introduces
performance inefficiency and metric noise. This behavior is reachable in common
scenarios involving monorepos with numerous workspaces.

_{Did we get this right? 👍 / 👎 to inform future reviews.}