fix(preprod): truncate max status check length (#102987)

Resolves EME-612

Author: trevor-e

src/sentry/preprod/vcs/status_checks/size/tasks.py

@@ -87,6 +87,7 @@ def create_preprod_status_check_task(preprod_artifact_id: int) -> None:
         client,
         commit_comparison.provider,
         preprod_artifact.project.organization_id,
+        preprod_artifact.project.organization.slug,
         repository.integration_id,
     )
     if not provider:
@@ -247,10 +248,16 @@ def _get_status_check_client(
 
 
 def _get_status_check_provider(
-    client: StatusCheckClient, provider: str | None, organization_id: int, integration_id: int
+    client: StatusCheckClient,
+    provider: str | None,
+    organization_id: int,
+    organization_slug: str,
+    integration_id: int,
 ) -> _StatusCheckProvider | None:
     if provider == IntegrationProviderSlug.GITHUB:
-        return _GitHubStatusCheckProvider(client, provider, organization_id, integration_id)
+        return _GitHubStatusCheckProvider(
+            client, provider, organization_id, organization_slug, integration_id
+        )
     else:
         return None
 
@@ -266,11 +273,13 @@ def __init__(
         client: StatusCheckClient,
         provider_key: str,
         organization_id: int,
+        organization_slug: str,
         integration_id: int,
     ):
         self.client = client
         self.provider_key = provider_key
         self.organization_id = organization_id
+        self.organization_slug = organization_slug
         self.integration_id = integration_id
 
     def _create_scm_interaction_event(self) -> SCMIntegrationInteractionEvent:
@@ -322,19 +331,44 @@ def create_status_check(
                 )
                 return None
 
+            truncated_text = _truncate_to_byte_limit(text, GITHUB_MAX_TEXT_FIELD_LENGTH)
+            truncated_summary = _truncate_to_byte_limit(summary, GITHUB_MAX_SUMMARY_FIELD_LENGTH)
+
+            if text and truncated_text and len(truncated_text) != len(text):
+                logger.warning(
+                    "preprod.status_checks.create.text_truncated",
+                    extra={
+                        "original_bytes": len(text.encode("utf-8")),
+                        "truncated_bytes": len(truncated_text.encode("utf-8")),
+                        "organization_id": self.organization_id,
+                        "organization_slug": self.organization_slug,
+                    },
+                )
+
+            if summary and truncated_summary and len(truncated_summary) != len(summary):
+                logger.warning(
+                    "preprod.status_checks.create.summary_truncated",
+                    extra={
+                        "original_bytes": len(summary.encode("utf-8")),
+                        "truncated_bytes": len(truncated_summary.encode("utf-8")),
+                        "organization_id": self.organization_id,
+                        "organization_slug": self.organization_slug,
+                    },
+                )
+
             check_data: dict[str, Any] = {
                 "name": title,
                 "head_sha": sha,
                 "external_id": external_id,
                 "output": {
                     "title": subtitle,
-                    "summary": summary,
+                    "summary": truncated_summary,
                 },
                 "status": mapped_status.value,
             }
 
-            if text:
-                check_data["output"]["text"] = text
+            if truncated_text:
+                check_data["output"]["text"] = truncated_text
 
             if mapped_conclusion:
                 check_data["conclusion"] = mapped_conclusion.value
@@ -395,6 +429,35 @@ def create_status_check(
                 raise
 
 
+# See: https://docs.github.com/en/rest/checks/runs?apiVersion=2022-11-28#create-a-check-run
+GITHUB_MAX_SUMMARY_FIELD_LENGTH = 65535
+GITHUB_MAX_TEXT_FIELD_LENGTH = 65535
+
+
+def _truncate_to_byte_limit(text: str | None, byte_limit: int) -> str | None:
+    """Truncate text to fit within byte limit while ensuring valid UTF-8."""
+    if not text:
+        return text
+
+    TRUNCATE_AMOUNT = 10
+
+    encoded = text.encode("utf-8")
+    if len(encoded) <= byte_limit:
+        return text
+
+    if byte_limit <= TRUNCATE_AMOUNT:
+        # This shouldn't happen, but just in case.
+        truncated = encoded[:byte_limit].decode("utf-8", errors="ignore")
+        return truncated
+
+    # Truncate to byte_limit - 10 (a bit of wiggle room) to make room for "..."
+    # Note: this can break formatting you have and is more of a catch-all,
+    # broken formatting is better than silently erroring for the user.
+    # Templating logic itself should try to more contextually trim the content if possible.
+    truncated = encoded[: byte_limit - TRUNCATE_AMOUNT].decode("utf-8", errors="ignore")
+    return truncated + "..."
+
+
 GITHUB_STATUS_CHECK_STATUS_MAPPING: dict[StatusCheckStatus, GitHubCheckStatus] = {
     StatusCheckStatus.ACTION_REQUIRED: GitHubCheckStatus.COMPLETED,
     StatusCheckStatus.IN_PROGRESS: GitHubCheckStatus.IN_PROGRESS,

tests/sentry/preprod/vcs/status_checks/size/test_status_checks_tasks.py

@@ -13,6 +13,7 @@
 from sentry.shared_integrations.exceptions import IntegrationConfigurationError
 from sentry.testutils.cases import TestCase
 from sentry.testutils.silo import region_silo_test
+from sentry.utils import json
 
 
 @region_silo_test
@@ -663,3 +664,66 @@ def test_create_preprod_status_check_task_github_429_error(self):
                 assert "rate limit" in str(e).lower()
 
         assert len(responses.calls) == 1
+
+    @responses.activate
+    def test_create_preprod_status_check_task_truncates_long_summary(self):
+        """Test task truncates summary when it exceeds GitHub's byte limit."""
+        commit_comparison = CommitComparison.objects.create(
+            organization_id=self.organization.id,
+            head_sha="a" * 40,
+            base_sha="b" * 40,
+            provider="github",
+            head_repo_name="owner/repo",
+            base_repo_name="owner/repo",
+            head_ref="feature/test",
+            base_ref="main",
+        )
+
+        artifacts = []
+        for i in range(150):
+            long_app_id = f"com.example.very.long.app.identifier.number.{i}" + "x" * 200
+            artifact = PreprodArtifact.objects.create(
+                project=self.project,
+                state=PreprodArtifact.ArtifactState.FAILED,
+                app_id=long_app_id,
+                error_message=f"This is a very long error message that will contribute to the summary size. Error #{i}: "
+                + "y" * 500,
+                commit_comparison=commit_comparison,
+            )
+            artifacts.append(artifact)
+
+        integration = self.create_integration(
+            organization=self.organization,
+            external_id="test-truncation",
+            provider="github",
+            metadata={"access_token": "test_token", "expires_at": "2099-01-01T00:00:00Z"},
+        )
+
+        Repository.objects.create(
+            organization_id=self.organization.id,
+            name="owner/repo",
+            provider="integrations:github",
+            integration_id=integration.id,
+        )
+
+        responses.add(
+            responses.POST,
+            "https://api.github.com/repos/owner/repo/check-runs",
+            status=201,
+            json={"id": 12345, "status": "completed"},
+        )
+
+        with self.tasks():
+            create_preprod_status_check_task(artifacts[0].id)
+
+        assert len(responses.calls) == 1
+        request_body = responses.calls[0].request.body
+
+        payload = json.loads(request_body)
+        summary = payload["output"]["summary"]
+
+        assert summary is not None
+        summary_bytes = len(summary.encode("utf-8"))
+
+        assert summary_bytes <= 65535, f"Summary has {summary_bytes} bytes, exceeds GitHub limit"
+        assert summary.endswith("..."), "Truncated summary should end with '...'"