Reapply "feat(repos): Add audit logs when auto syncing repositories (#112056)"

This reverts commit 2765803.

Author: wedamija

src/sentry/audit_log/events.py

@@ -384,3 +384,38 @@ def render(self, audit_log_entry: AuditLogEntry) -> str:
         return "updated repository settings for {repository_count} repositories".format(
             repository_count=data.get("repository_count", 0),
         )
+
+
+def _render_repo_event(action: str, audit_log_entry: AuditLogEntry) -> str:
+    data = audit_log_entry.data
+    actor = audit_log_entry.actor_label or "unknown"
+    repo_name = data.get("repo_name", "unknown")
+    source = data.get("source", "")
+    msg = f"{actor} {action} repository {repo_name}"
+    if source:
+        msg += f" (via {source})"
+    return msg
+
+
+class RepoAddedAuditLogEvent(AuditLogEvent):
+    def __init__(self) -> None:
+        super().__init__(event_id=1161, name="REPO_ADDED", api_name="repo.added")
+
+    def render(self, audit_log_entry: AuditLogEntry) -> str:
+        return _render_repo_event("added", audit_log_entry)
+
+
+class RepoDisabledAuditLogEvent(AuditLogEvent):
+    def __init__(self) -> None:
+        super().__init__(event_id=1162, name="REPO_DISABLED", api_name="repo.disabled")
+
+    def render(self, audit_log_entry: AuditLogEntry) -> str:
+        return _render_repo_event("disabled", audit_log_entry)
+
+
+class RepoEnabledAuditLogEvent(AuditLogEvent):
+    def __init__(self) -> None:
+        super().__init__(event_id=1163, name="REPO_ENABLED", api_name="repo.enabled")
+
+    def render(self, audit_log_entry: AuditLogEntry) -> str:
+        return _render_repo_event("enabled", audit_log_entry)

src/sentry/audit_log/register.py

@@ -695,3 +695,6 @@
         template="updated autofix automation settings for {project_count} projects",
     )
 )
+default_manager.add(events.RepoAddedAuditLogEvent())
+default_manager.add(events.RepoDisabledAuditLogEvent())
+default_manager.add(events.RepoEnabledAuditLogEvent())

src/sentry/integrations/github/tasks/link_all_repos.py

@@ -13,7 +13,6 @@
 )
 from sentry.organizations.services.organization import organization_service
 from sentry.plugins.providers.integration_repository import (
-    RepoExistsError,
     RepositoryInputConfig,
     get_integration_repository_provider,
 )
@@ -40,7 +39,7 @@ def get_repo_config(repo: Mapping[str, Any], integration_id: int) -> RepositoryI
     processing_deadline_duration=60,
     silo_mode=SiloMode.CONTROL,
 )
-@retry(exclude=(RepoExistsError, KeyError))
+@retry(exclude=(KeyError,))
 def link_all_repos(
     integration_key: str,
     integration_id: int,
@@ -88,14 +87,15 @@ def link_all_repos(
                 missing_repos.append(repo)
                 continue
 
-        try:
+        _created_repos, _reactivated_repos, existing_repos = (
             integration_repo_provider.create_repositories(
                 configs=repo_configs, organization=rpc_org
             )
-        except RepoExistsError as e:
+        )
+        if existing_repos:
             lifecycle.record_halt(
                 str(LinkAllReposHaltReason.REPOSITORY_NOT_CREATED),
-                {"missing_repos": e.repos, "integration_id": integration_id},
+                {"missing_repos": existing_repos, "integration_id": integration_id},
             )
 
         if missing_repos:

src/sentry/integrations/github/tasks/sync_repos.py

@@ -21,11 +21,9 @@
     SCMIntegrationInteractionEvent,
     SCMIntegrationInteractionType,
 )
+from sentry.integrations.source_code_management.repo_audit import log_repo_change
 from sentry.organizations.services.organization import organization_service
-from sentry.plugins.providers.integration_repository import (
-    RepoExistsError,
-    get_integration_repository_provider,
-)
+from sentry.plugins.providers.integration_repository import get_integration_repository_provider
 from sentry.shared_integrations.exceptions import ApiError
 from sentry.silo.base import SiloMode
 from sentry.tasks.base import instrumented_task, retry
@@ -171,6 +169,8 @@ def sync_repos_for_org(organization_integration_id: int) -> None:
         if dry_run:
             return
 
+        repo_by_external_id = {r.external_id: r for r in active_repos + disabled_repos}
+
         if new_ids:
             integration_repo_provider = get_integration_repository_provider(integration)
             repo_configs = [
@@ -179,12 +179,27 @@ def sync_repos_for_org(organization_integration_id: int) -> None:
                 if str(repo["id"]) in new_ids
             ]
             if repo_configs:
-                try:
-                    integration_repo_provider.create_repositories(
-                        configs=repo_configs, organization=rpc_org
+                created_repos, reactivated_repos, _ = integration_repo_provider.create_repositories(
+                    configs=repo_configs, organization=rpc_org
+                )
+
+                for repo in created_repos:
+                    log_repo_change(
+                        event_name="REPO_ADDED",
+                        organization_id=organization_id,
+                        repo=repo,
+                        source="repository sync",
+                        provider=integration.provider,
+                    )
+
+                for repo in reactivated_repos:
+                    log_repo_change(
+                        event_name="REPO_ENABLED",
+                        organization_id=organization_id,
+                        repo=repo,
+                        source="repository sync",
+                        provider=integration.provider,
                     )
-                except RepoExistsError:
-                    pass
 
         if removed_ids:
             repository_service.disable_repositories_by_external_ids(
@@ -194,13 +209,31 @@ def sync_repos_for_org(organization_integration_id: int) -> None:
                 external_ids=list(removed_ids),
             )
 
+            for eid in removed_ids:
+                removed_repo = repo_by_external_id.get(eid)
+                if removed_repo:
+                    log_repo_change(
+                        event_name="REPO_DISABLED",
+                        organization_id=organization_id,
+                        repo=removed_repo,
+                        source="automatic SCM syncing",
+                        provider=integration.provider,
+                    )
+
         if restored_ids:
             for repo in disabled_repos:
                 if repo.external_id in restored_ids:
                     repo.status = ObjectStatus.ACTIVE
                     repository_service.update_repository(
                         organization_id=organization_id, update=repo
                     )
+                    log_repo_change(
+                        event_name="REPO_ENABLED",
+                        organization_id=organization_id,
+                        repo=repo,
+                        source="automatic SCM syncing",
+                        provider=integration.provider,
+                    )
 
 
 @instrumented_task(

src/sentry/integrations/github/tasks/sync_repos_on_install_change.py

@@ -13,10 +13,10 @@
     SCMIntegrationInteractionEvent,
     SCMIntegrationInteractionType,
 )
+from sentry.integrations.source_code_management.repo_audit import log_repo_change
 from sentry.organizations.services.organization import organization_service
 from sentry.organizations.services.organization.model import RpcOrganization
 from sentry.plugins.providers.integration_repository import (
-    RepoExistsError,
     RepositoryInputConfig,
     get_integration_repository_provider,
 )
@@ -36,7 +36,7 @@
     processing_deadline_duration=120,
     silo_mode=SiloMode.CONTROL,
 )
-@retry(exclude=(RepoExistsError, KeyError))
+@retry(exclude=(KeyError,))
 def sync_repos_on_install_change(
     integration_id: int,
     action: str,
@@ -119,18 +119,59 @@ def _sync_repos_for_org(
                 continue
 
         if repo_configs:
-            try:
+            created_repos, reactivated_repos, _missing_repos = (
                 integration_repo_provider.create_repositories(
                     configs=repo_configs, organization=rpc_org
                 )
-            except RepoExistsError:
-                pass
+            )
+
+            for created_repo in created_repos:
+                log_repo_change(
+                    event_name="REPO_ADDED",
+                    organization_id=rpc_org.id,
+                    repo=created_repo,
+                    source="GitHub webhook",
+                    provider=integration.provider,
+                )
+
+            for reactivated_repo in reactivated_repos:
+                log_repo_change(
+                    event_name="REPO_ENABLED",
+                    organization_id=rpc_org.id,
+                    repo=reactivated_repo,
+                    source="GitHub webhook",
+                    provider=integration.provider,
+                )
 
     if repos_removed:
+        # Look up repos before disabling to get their IDs and names
         external_ids = [str(repo["id"]) for repo in repos_removed]
+        existing_repos = repository_service.get_repositories(
+            organization_id=rpc_org.id,
+            integration_id=integration.id,
+            providers=[provider],
+        )
+        repo_by_eid = {
+            r.external_id: r
+            for r in existing_repos
+            if r.external_id and r.status == ObjectStatus.ACTIVE
+        }
+
         repository_service.disable_repositories_by_external_ids(
             organization_id=rpc_org.id,
             integration_id=integration.id,
             provider=provider,
             external_ids=external_ids,
         )
+
+        for repo in repos_removed:
+            eid = str(repo["id"])
+            sentry_repo = repo_by_eid.get(eid)
+            if sentry_repo:
+                log_repo_change(
+                    event_name="REPO_DISABLED",
+                    organization_id=rpc_org.id,
+                    repo=sentry_repo,
+                    source="GitHub webhook",
+                    provider=integration.provider,
+                )

src/sentry/integrations/source_code_management/repo_audit.py

@@ -0,0 +1,23 @@
+"""
+Audit log helpers for repository sync operations.
+"""
+
+from sentry import audit_log
+from sentry.integrations.services.repository.model import RpcRepository
+from sentry.utils.audit import create_system_audit_entry
+
+
+def log_repo_change(
+    *, event_name: str, organization_id: int, repo: RpcRepository, source: str, provider: str
+) -> None:
+    create_system_audit_entry(
+        organization_id=organization_id,
+        target_object=repo.id,
+        event=audit_log.get_event_id(event_name),
+        data={
+            "repo_name": repo.name,
+            "external_id": repo.external_id,
+            "source": source,
+            "provider": provider,
+        },
+    )

src/sentry/plugins/providers/integration_repository.py

@@ -240,13 +240,21 @@ def create_repositories(
         self,
         configs: list[RepositoryInputConfig],
         organization: RpcOrganization,
-    ):
+    ) -> tuple[list[RpcRepository], list[RpcRepository], list[RepositoryConfig]]:
+        """
+        Create or update repositories from configs.
+        Returns (created, reactivated, missing) — newly created repos, repos that
+        were reactivated or updated from a hidden/unlinked state, and repo configs
+        that could not be created because a repository with that configuration
+        already exists.
+        """
         external_id_to_repo_config: dict[str, RepositoryConfig] = {}
         for config in configs:
             result = self.build_repository_config(organization=organization, data=config)
             external_id_to_repo_config[result["external_id"]] = result
 
         repos_to_update: list[RpcRepository] = []
+        created_repos: list[RpcRepository] = []
 
         hidden_repos = repository_service.get_repositories(
             organization_id=organization.id,
@@ -272,6 +280,7 @@ def create_repositories(
                 organization_id=organization.id, create=create_repository
             )
             if new_repository is not None:
+                created_repos.append(new_repository)
                 continue
 
             missing_repos.append(repo_config)
@@ -299,8 +308,7 @@ def create_repositories(
                 updates=repos_to_update,
             )
 
-        if missing_repos:
-            raise RepoExistsError(repos=missing_repos)
+        return created_repos, repos_to_update, missing_repos
 
     def dispatch(self, request: Request, organization, **kwargs):
         try:

tests/sentry/integrations/github/tasks/test_sync_repos.py

@@ -4,14 +4,16 @@
 import responses
 from taskbroker_client.retry import RetryTaskError
 
+from sentry import audit_log
 from sentry.constants import ObjectStatus
 from sentry.integrations.github.integration import GitHubIntegrationProvider
 from sentry.integrations.github.tasks.sync_repos import sync_repos_for_org
 from sentry.integrations.models.organization_integration import OrganizationIntegration
+from sentry.models.auditlogentry import AuditLogEntry
 from sentry.models.repository import Repository
 from sentry.silo.base import SiloMode
 from sentry.testutils.cases import IntegrationTestCase
-from sentry.testutils.silo import assume_test_silo_mode, control_silo_test
+from sentry.testutils.silo import assume_test_silo_mode, assume_test_silo_mode_of, control_silo_test
 
 
 @control_silo_test
@@ -60,6 +62,13 @@ def test_creates_new_repos(self, _: MagicMock) -> None:
         assert repos[0].provider == "integrations:github"
         assert repos[1].name == "getsentry/snuba"
 
+        with assume_test_silo_mode_of(AuditLogEntry):
+            entries = AuditLogEntry.objects.filter(
+                organization_id=self.organization.id,
+                event=audit_log.get_event_id("REPO_ADDED"),
+            )
+            assert entries.count() == 2
+
     @responses.activate
     def test_disables_removed_repos(self, _: MagicMock) -> None:
         with assume_test_silo_mode(SiloMode.CELL):
@@ -89,6 +98,16 @@ def test_disables_removed_repos(self, _: MagicMock) -> None:
                 organization_id=self.organization.id, external_id="1"
             ).exists()
 
+        with assume_test_silo_mode_of(AuditLogEntry):
+            assert AuditLogEntry.objects.filter(
+                organization_id=self.organization.id,
+                event=audit_log.get_event_id("REPO_DISABLED"),
+            ).exists()
+            assert AuditLogEntry.objects.filter(
+                organization_id=self.organization.id,
+                event=audit_log.get_event_id("REPO_ADDED"),
+            ).exists()
+
     @responses.activate
     def test_re_enables_restored_repos(self, _: MagicMock) -> None:
         with assume_test_silo_mode(SiloMode.CELL):
@@ -113,6 +132,12 @@ def test_re_enables_restored_repos(self, _: MagicMock) -> None:
             repo.refresh_from_db()
             assert repo.status == ObjectStatus.ACTIVE
 
+        with assume_test_silo_mode_of(AuditLogEntry):
+            assert AuditLogEntry.objects.filter(
+                organization_id=self.organization.id,
+                event=audit_log.get_event_id("REPO_ENABLED"),
+            ).exists()
+
     @responses.activate
     def test_no_changes_needed(self, _: MagicMock) -> None:
         with assume_test_silo_mode(SiloMode.CELL):