feat(preprod): Add snapshot image comparison task and endpoint logic

[NicoHinderling]

Adds the ability to automatically compare snapshot images between two preprod builds using pixel-level diffing.

When a snapshot is uploaded with VCS info (base_sha, base_repo_name, base_ref), the endpoint now looks for a matching base artifact and triggers a compare_snapshots Celery task. This task:

• Fetches manifests for both head and base artifacts
• Matches images by file name across the two builds
• Runs pixel-level comparison using odiff (via a new image_diff module)
• Stores diff masks and a comparison manifest in object storage
• Records results in a PreprodSnapshotComparison model

The GET endpoint is updated to return categorized image lists (changed, added, removed, unchanged) along with diff metadata when comparison data is available.

New modules:

• sentry.preprod.snapshots.image_diff — wraps the odiff binary in server mode for efficient batch image comparison
• sentry.preprod.snapshots.tasks — Celery task for async snapshot comparison
• snapshot-diff internal endpoint registered in URL patterns

Also adds odiff-bin as a dependency for the image diffing binary.

[github-actions]

🚨 Warning: This pull request contains Frontend and Backend changes!

It's discouraged to make changes to Sentry's Frontend and Backend in a single pull request. The Frontend and Backend are not atomically deployed. If the changes are interdependent of each other, they must be separated into two pull requests and be made forward or backwards compatible, such that the Backend or Frontend can be safely deployed independently.

Have questions? Please ask in the #discuss-dev-infra channel.

[NicoHinderling]

• feat(preprod): Add snapshot image comparison task and endpoint logic #109151 : 3 dependent PRs (#109397 , #109403 , #109546 ) 👈 (View in Graphite)
• feat(preprod): Add snapshot comparison task #109382
• feat(preprod): Add image comparison library with odiff batch support #109381
• feat(preprod): Add odiff server wrapper and Dockerfile binary install #109380
• master

This stack of pull requests is managed by Graphite. Learn more about stacking.

[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.}

[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.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

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

[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.}

[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.}

[cursor]

Errored items missing available dimension data in manifest

Medium Severity

For exceeds_pixel_limit errored items, the task has head_meta and base_meta with width/height readily available (lines 241-242) but doesn't store them in image_results. The "removed" items were fixed to include before_width/before_height per PR discussion, but errored items were missed. In the categorizer fallback path (when head_img or base_img is None), img.before_width, img.after_width, etc. are all None, causing dimensions to silently fall back to 0. This produces incorrect dimension metadata for errored diff pairs in the API response.

Additional Locations (1)

• src/sentry/preprod/snapshots/comparison_categorizer.py#L82-L96

 

[NicoHinderling]

we don't need the right dimensions for error cases since they couldn't be processed and therefore can't be displayed