feat(preprod): Add Size Analysis detector

[chromy]

Add PreprodSizeAnalysisGroupType (type_id=11003) with detector handler
and validators that integrate with the workflow engine to emit issues
when build size thresholds are exceeded.

• PreprodSizeAnalysisDetectorHandler: evaluates DataPackets from size
  comparisons and produces IssueOccurrences
• PreprodSizeAnalysisDetectorValidato validate detector creation via the API
• Register the GroupType import in preprod/grouptype.py

PRs:

• feat(preprod): Add Size Analysis detector #108208 Add size_analysis detector (this PR)
• feat(preprod): Hook size analysis detector to diff #108209 Hook size analysis detector to diff
• feat(preprod): Add new issue type to frontend #108210 Add new issue type to frontend
• feat(preprod): Add size monitor UI #108211 Add size monitor UI

Design doc

This stack of PRs is missing a handful of important features:

• Filters
• Detailed data on the occurrence
• Some UI polish

but it works end to end.

[github-actions]

This PR has a migration; here is the generated SQL for src/sentry/preprod/migrations/0027_size_analysis_subscription.py

for 0027_size_analysis_subscription in preprod

--
-- Create model SizeAnalysisSubscription
--
CREATE TABLE "sentry_sizeanalysissubscription" ("id" bigint NOT NULL PRIMARY KEY GENERATED BY DEFAULT AS IDENTITY, "date_updated" timestamp with time zone NOT NULL, "date_added" timestamp with time zone NOT NULL, "project_id" bigint NOT NULL);
ALTER TABLE "sentry_sizeanalysissubscription" ADD CONSTRAINT "sentry_sizeanalysiss_project_id_41e3355e_fk_sentry_pr" FOREIGN KEY ("project_id") REFERENCES "sentry_project" ("id") DEFERRABLE INITIALLY DEFERRED NOT VALID;
ALTER TABLE "sentry_sizeanalysissubscription" VALIDATE CONSTRAINT "sentry_sizeanalysiss_project_id_41e3355e_fk_sentry_pr";
CREATE INDEX CONCURRENTLY "sentry_sizeanalysissubscription_project_id_41e3355e" ON "sentry_sizeanalysissubscription" ("project_id");

[saponifi3d]

the workflow engine code lgtm -- added a few notes for myself on abstractions that should probably live on the base class.

[saponifi3d]

🙏 -- this is more or less what the eventual base abstraction will look like too. i'll update this once we start working on the non-Stateful abstraction.

[saponifi3d]

note for myself -- this feels like another top level abstraction we should pull out as part of the ABC

[saponifi3d]

outside of evaluate_impl and _evaluate_conditions -- was there anything else you felt like should live in the base class? i can take a look at updating everything there.

[sentry]

Bug: The PreprodSizeAnalysisDetectorHandler hardcodes the platform as "android", which will cause incorrect platform attribution for issues generated from iOS/Apple builds.
_{Severity: MEDIUM}

Suggested Fix

Modify the data pipeline to include platform information in the DataPacket[SizeAnalysisValue] sent to the detector. This could involve adding a platform field to the SizeAnalysisValue TypedDict or looking up the artifact type from an identifier within the packet. The handler should then use this dynamic platform value instead of the hardcoded "android" string.

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/preprod/size_analysis/grouptype.py#L158

Potential issue: The `create_occurrence` method in the new
`PreprodSizeAnalysisDetectorHandler` hardcodes the `platform` to "android". This is
incorrect because the platform should be determined dynamically based on the artifact
type, as is done in the existing `diff_to_occurrence` function. The data packet
(`SizeAnalysisValue`) passed to the handler does not currently contain the necessary
artifact type or platform information. If this detector is activated for iOS builds,
which use the `XCARCHIVE` artifact type, any generated issues will be incorrectly
attributed to the Android platform, impacting platform-specific filtering and metrics.

[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]

Division by zero when base size is zero

Medium Severity

In extract_value, the relative_diff case computes (head - base) / base. If base is 0 (e.g., a new measurement category that previously had no size), this raises a ZeroDivisionError. The _extract_base method only validates that the value is not None, so a zero value passes validation and reaches the division.