feat(preprod): Support absolute detectors from single-build path

[chromy]

Add maybe_emit_issues_from_size_results so that "absolute" threshold detectors fire as soon as a single build's size metrics are committed, without waiting for a comparison against a base artifact.

Previously, all issue emission went through the comparison path (maybe_emit_issues), which requires both head and base artifacts. This meant absolute detectors — which only need one build's size — couldn't fire until a comparison happened. Now:

• The single-build path (called after metrics commit in _assemble_preprod_artifact_size_analysis) processes only threshold_type == "absolute" detectors
• The comparison path (_maybe_emit_issues) processes only threshold_type in ("absolute_diff", "relative_diff") detectors
• SizeAnalysisMetadata base fields (base_metric_id, base_artifact_id, base_artifact) are now NotRequired, and create_occurrence conditionally populates base-related evidence and tags

Agent transcript: https://claudescope.sentry.dev/share/UgFhB6vtmF1RiOVM-Ble2aoN68w7YgfkO_BzAXfdxnE

[NicoHinderling]

it would be nice if the values here were in some array const or something in the same file as all of the options so that it's easy to maintain if we ever add more options

[chromy]

done, thanks!

[NicoHinderling]

since this is so specific to the "absolute" type of metrics, perhaps we should have the method title include that fact?

perhaps

_maybe_emit_issues_from_absolute_size_results or something?

[chromy]

done, thanks!

[cursor]

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

Bugbot Autofix prepared a fix for the issue found in the latest run.

• ✅ Fixed: Constants placed between import statements
  • Moved DIFF_THRESHOLD_TYPES and ABSOLUTE_THRESHOLD_TYPES constants to after the import block, following Python conventions.

Or push these changes by commenting:

@cursor push 41940c7851

Preview (41940c7851)

diff --git a/src/sentry/preprod/size_analysis/tasks.py b/src/sentry/preprod/size_analysis/tasks.py
--- a/src/sentry/preprod/size_analysis/tasks.py
+++ b/src/sentry/preprod/size_analysis/tasks.py
@@ -28,14 +28,14 @@
 from sentry.tasks.base import instrumented_task
 from sentry.taskworker.namespaces import preprod_tasks
 from sentry.utils import metrics
+from sentry.utils.json import dumps_htmlsafe
+from sentry.workflow_engine.models import DataPacket, Detector
+from sentry.workflow_engine.processors.detector import process_detectors
 
 # Threshold type categories for filtering detectors by path.
 # These must stay in sync with the enum in PreprodSizeAnalysisGroupType.detector_settings.config_schema.
 DIFF_THRESHOLD_TYPES = frozenset({"absolute_diff", "relative_diff"})
 ABSOLUTE_THRESHOLD_TYPES = frozenset({"absolute"})
-from sentry.utils.json import dumps_htmlsafe
-from sentry.workflow_engine.models import DataPacket, Detector
-from sentry.workflow_engine.processors.detector import process_detectors
 
 logger = logging.getLogger(__name__)

_{This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.}

[cursor]

Constants placed between import statements

Low Severity

DIFF_THRESHOLD_TYPES and ABSOLUTE_THRESHOLD_TYPES are defined in the middle of the import block, with from sentry.utils.json import dumps_htmlsafe immediately following on line 36. This appears to be an accidental placement — the constants split the imports, reducing readability and violating standard Python conventions for keeping all imports together at the top of a file. The constants belong after the import block finishes.

 

[sentry]

Bug: The function _maybe_emit_issues_from_absolute_size_results causes an N+1 query by accessing head_metric.preprod_artifact.project.organization in a loop without preloading these relations.
_{Severity: MEDIUM}

Suggested Fix

Preload the necessary relations when fetching the metrics. Add .select_related('preprod_artifact__project__organization') to the query that populates the metrics_in_transaction list before it is iterated over.

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/tasks.py#L656-L658

Potential issue: The function `_maybe_emit_issues_from_absolute_size_results` is called
within a loop for each metric. Inside the function, accessing
`head_metric.preprod_artifact.project.organization` triggers separate database queries
for `preprod_artifact`, `project`, and `organization` because these related objects are
not preloaded. This results in an N+1 query pattern, which can degrade the performance
of size analysis processing, particularly for builds that involve multiple components.

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