Use `resource.yml` in field classification in config-remote-sync by ilyakuz-db · Pull Request #4677 · databricks/cli

ilyakuz-db · 2026-03-06T16:37:37Z

Changes

bundle config-remote-sync derives its field filtering from the resource lifecycle metadata (resources.yml + resources.generated.yml) instead of the hand-maintained serverSideDefaults table:

ignore_remote_changes fields are excluded from sync unconditionally; backend_defaults fields only when absent from config, honoring values.
The metadata is matched directly, never via plan actions or OverrideChangeDesc hooks — those answer "what will deploy do", not "what belongs in config" (e.g. dashboard etag is re-promoted to Update for modified-remotely detection, yet must never be synced).
Sync-specific rules stay in the configsync package: jobs edit_mode, queue reset values, name-prefix stripping, and a node_type_id keep-list so remotely added job clusters stay deployable.
The matching helpers move unchanged to dresources/match.go so plan and sync share one implementation.

resources.yml changes:

custom_tags / cluster_log_conf become backend_defaults on job and standalone clusters (moved from the configsync-only rule in configsync: don't sync policy-injected custom_tags and cluster_log_conf #5610, per its TODO). This also stops perpetual direct-engine plan updates for policy-injected values.
New backend defaults: performance_target ["PERFORMANCE_OPTIMIZED"], task timeout_seconds [0] / disabled [false] / pipeline_task.full_refresh [false], email_notifications.no_alert_for_skipped_runs [false], task/job-cluster single_user_name, cluster driver/worker_node_type_flexibility; dashboards etag → ignore_remote_changes (inert for deploy). The zero-value rules are inert for the plan (the all-empty check runs before backend defaults, and any other remote value fails the values constraint); they exist for config-remote-sync's nested entity filtering.
jobs/clusters node_type_id entries are kept — policy- and pool-backed clusters materialize it remotely while config legitimately omits it. Only the never-matching pipelines clusters[*].node_type_id entry is removed (pipelines/get returns the stored spec verbatim).

Two intentional sync semantic changes, pinned by unit tests: (a) a remote value that filters entirely to defaults emits remove instead of replace {}; (b) backend_defaults without values strip any remote value when the field is absent from config (the old table matched one specific value). One narrow divergence: a remote-only data_security_mode on a standalone cluster now syncs verbatim instead of being stripped when equal to SINGLE_USER — its resources.yml rule is deliberately absent because clusters use the alias handler in cluster.go.

Why

The table constantly lagged behind resources.yml: every new field had to be added in two places, and gaps leaked server-side values into users' YAML. Lifecycle metadata added for the deploy plan now automatically protects config-remote-sync.

Tests

Unit tests for the metadata-driven skip logic and nested entity filtering; all related acceptance suites pass on both engines with no output changes.
New acceptance test pins the plan behavior change: policy-injected custom_tags/cluster_log_conf alone produce no update; a real remote edit still does.
Live state/local/remote presence matrix on a real AWS workspace: for every touched field, plan classification differs from v1.5.0 only in the (absent from state, absent from config, present remotely) cell; in particular, removing a previously-deployed field from config still plans an update on both versions.
Live A/B on a real AWS workspace (v1.5.0 vs this branch, direct engine, policy fixing node_type_id/custom_tags/cluster_log_conf/spark_conf): spurious plan updates drop to spark_conf only (known gap, unchanged), no new drift fields, legit remote edits detected and synced identically, deploy end state identical; changing the policy's fixed node type after deploy stays suppressed on both.

eng-dev-ecosystem-bot · 2026-03-06T16:49:42Z

Integration test report

Commit: f7b3de5

Run: 28596848662

	Env	❌FAIL	🟨KNOWN	💚RECOVERED	🙈SKIP	✅pass	🙈skip	Time
💚	aws linux			10	13	230	1042	4:40
💚	aws windows			10	13	232	1040	3:52
💚	aws-ucws linux			10	13	314	960	5:19
💚	aws-ucws windows			10	13	316	958	4:25
💚	azure linux			4	15	230	1041	4:25
💚	azure windows			4	15	232	1039	3:35
💚	azure-ucws linux			4	15	316	957	6:05
💚	azure-ucws windows			4	15	318	955	4:02
❌	gcp linux	201	2		15	28	1032	1:50
❌	gcp windows	203	2		15	28	1030	2:44

226 interesting tests: 203 FAIL, 13 SKIP, 8 RECOVERED, 2 KNOWN

	Test Name	aws linux	aws windows	aws-ucws linux	aws-ucws windows	azure linux	azure windows	azure-ucws linux	azure-ucws windows	gcp linux	gcp windows
🟨	TestAccept	💚R	💚R	💚R	💚R	💚R	💚R	💚R	💚R	🟨K	🟨K
❌	TestAccept/bundle/deploy/mlops-stacks	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/bundle/deploy/mlops-stacks/DATABRICKS_BUNDLE_ENGINE=direct	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/bundle/deploy/mlops-stacks/DATABRICKS_BUNDLE_ENGINE=terraform	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
🙈	TestAccept/bundle/invariant/no_drift	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/permissions	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
💚	TestAccept/bundle/resources/permissions/jobs/destroy_without_mgmtperms/with_permissions	💚R	💚R	💚R	💚R	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
💚	TestAccept/bundle/resources/permissions/jobs/destroy_without_mgmtperms/with_permissions/DATABRICKS_BUNDLE_ENGINE=direct	💚R	💚R	💚R	💚R
💚	TestAccept/bundle/resources/permissions/jobs/destroy_without_mgmtperms/with_permissions/DATABRICKS_BUNDLE_ENGINE=terraform	💚R	💚R	💚R	💚R
💚	TestAccept/bundle/resources/permissions/jobs/destroy_without_mgmtperms/without_permissions	💚R	💚R	💚R	💚R	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
💚	TestAccept/bundle/resources/permissions/jobs/destroy_without_mgmtperms/without_permissions/DATABRICKS_BUNDLE_ENGINE=direct	💚R	💚R	💚R	💚R
💚	TestAccept/bundle/resources/permissions/jobs/destroy_without_mgmtperms/without_permissions/DATABRICKS_BUNDLE_ENGINE=terraform	💚R	💚R	💚R	💚R
🙈	TestAccept/bundle/resources/postgres_branches/basic	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/postgres_branches/recreate	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/postgres_branches/replace_existing	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/postgres_branches/update_protected	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/postgres_branches/without_branch_id	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/postgres_endpoints/basic	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/postgres_projects/update_display_name	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/synced_database_tables/basic	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/vector_search_endpoints/drift/recreated_same_name	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
🙈	TestAccept/bundle/resources/vector_search_indexes/recreate/embedding_dimension	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
❌	TestAccept/selftest/record_cloud/basic	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/basic/DATABRICKS_BUNDLE_ENGINE=direct	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/basic/DATABRICKS_BUNDLE_ENGINE=terraform	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/error	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/error/DATABRICKS_BUNDLE_ENGINE=direct	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/error/DATABRICKS_BUNDLE_ENGINE=terraform	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/pipeline-crud	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/pipeline-crud/DATABRICKS_BUNDLE_ENGINE=direct	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/pipeline-crud/DATABRICKS_BUNDLE_ENGINE=terraform	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/workspace-file-io	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/workspace-file-io/DATABRICKS_BUNDLE_ENGINE=direct	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAccept/selftest/record_cloud/workspace-file-io/DATABRICKS_BUNDLE_ENGINE=terraform	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
🙈	TestAccept/ssh/connection	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S	🙈S
❌	TestDashboardAssumptions_WorkspaceImport	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestBundleInitErrorOnUnknownFields	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestBundleInitHelpers	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAlertsCreateErrWhenNoArguments	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestApiGet	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestAuthDescribeSuccess	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestClustersGet	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestClustersList	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCat	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCat/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCatDoesNotSupportOutputModeJson	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCatForDbfsInvalidScheme	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCatOnADir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCatOnADir/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCatOnNonExistentFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCatOnNonExistentFile/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCompletion	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDir/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDir/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDir/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpDir/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDir/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirFileNotOverwritten	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirFileNotOverwritten/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirFileNotOverwritten/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirFileNotOverwritten/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirFileNotOverwritten/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirFileNotOverwritten/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirWithOverwriteFlag	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirWithOverwriteFlag/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirWithOverwriteFlag/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirWithOverwriteFlag/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirWithOverwriteFlag/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpDirToDirWithOverwriteFlag/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpErrorsOnInvalidScheme	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpErrorsWhenSourceIsDirWithoutRecursiveFlag	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpErrorsWhenSourceIsDirWithoutRecursiveFlag/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDir/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDir/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDir/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDir/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDir/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirFileNotOverwritten	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirFileNotOverwritten/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirFileNotOverwritten/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirFileNotOverwritten/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirFileNotOverwritten/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirFileNotOverwritten/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirForWindowsPaths	🙈s	✅p	🙈s	✅p	🙈s	✅p	🙈s	✅p	🙈s	❌F
❌	TestFsCpFileToDirWithOverwriteFlag	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirWithOverwriteFlag/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirWithOverwriteFlag/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirWithOverwriteFlag/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirWithOverwriteFlag/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToDirWithOverwriteFlag/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFile/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFile/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFile/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFile/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFile/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileFileNotOverwritten	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileFileNotOverwritten/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileFileNotOverwritten/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileFileNotOverwritten/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileFileNotOverwritten/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileFileNotOverwritten/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileWithOverwriteFlag	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileWithOverwriteFlag/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileWithOverwriteFlag/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileWithOverwriteFlag/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileWithOverwriteFlag/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToFileWithOverwriteFlag/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDir/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDir/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDir/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDir/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDir/local_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpFileToNonExistentDirWindowsPaths	🙈s	✅p	🙈s	✅p	🙈s	✅p	🙈s	✅p	🙈s	❌F
❌	TestFsCpSourceIsDirectoryButTargetIsFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpSourceIsDirectoryButTargetIsFile/dbfs_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpSourceIsDirectoryButTargetIsFile/dbfs_to_local	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsCpSourceIsDirectoryButTargetIsFile/dbfs_to_uc-volumes	🙈s	🙈s	✅p	✅p	🙈s	🙈s	✅p	✅p	❌F	❌F
❌	TestFsCpSourceIsDirectoryButTargetIsFile/local_to_dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLs/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsForNonexistingDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsForNonexistingDir/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsOnEmptyDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsOnEmptyDir/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsOnFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsOnFile/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsWithAbsolutePaths	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsWithAbsolutePaths/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsLsWithoutScheme	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdir/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdirCreatesIntermediateDirectories	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdirCreatesIntermediateDirectories/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdirWhenDirectoryAlreadyExists	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdirWhenDirectoryAlreadyExists/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdirWhenFileExistsAtPath	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsMkdirWhenFileExistsAtPath/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmDirRecursively	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmDirRecursively/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmEmptyDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmEmptyDir/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmFile/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmForNonExistentFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmForNonExistentFile/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmNonEmptyDirectory	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestFsRmNonEmptyDirectory/dbfs	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSecretsPutSecretBytesValue	✅p	✅p	🙈s	🙈s	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSecretsPutSecretStringValue	✅p	✅p	🙈s	🙈s	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncEnsureRemotePathIsUsableIfRepoDoesntExist	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncEnsureRemotePathIsUsableIfRepoExists	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncEnsureRemotePathIsUsableInWorkspace	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncFullFileSync	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncIncrementalFileOverwritesFolder	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncIncrementalFileSync	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncIncrementalSyncFileToPythonNotebook	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncIncrementalSyncPythonNotebookDelete	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncIncrementalSyncPythonNotebookToFile	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncNestedFolderDoesntFailOnNonEmptyDirectory	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncNestedFolderSync	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestSyncNestedSpacePlusAndHashAreEscapedSync	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestExport	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestExportDir	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestExportDirDoesNotOverwrite	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestExportDirWithOverwriteFlag	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F
❌	TestExportWithFileFlag	✅p	✅p	✅p	✅p	✅p	✅p	✅p	✅p	❌F	❌F

Top 4 slowest tests (at least 2 minutes):

duration	env	testname
3:06	aws-ucws windows	TestAccept
2:45	azure windows	TestAccept
2:39	aws windows	TestAccept
2:38	azure-ucws windows	TestAccept

(56 table rows omitted to keep the report under 60000 bytes)

ilyakuz-db · 2026-03-07T12:32:07Z

  email_notifications.on_failure[0]: replace
  max_concurrent_runs: replace
  tags['env']: remove
+  timeout_seconds: remove


This is expected, logic was incorrect previously - this case was skipped completely

ilyakuz-db · 2026-03-07T12:32:42Z

+// If a CLI-defaulted field is changed on remote and should be disabled
+// (e.g. queueing disabled → remote field is nil), we can't define it
+// in the config as "null" because the CLI default will be applied again.
+var resetValues = map[string][]resetRule{


Should we include this logic in resources.yml?

ilyakuz-db · 2026-03-07T12:33:58Z

        reason: managed

    backend_defaults:
+      # Backend returns PERFORMANCE_OPTIMIZED for serverless jobs when not set


This is set only for serverless jobs, probably, this should live in changes desc overrides

Is this field ever set for non-serverless jobs?

It is, actually! I thought that it only applies to Serverless jobs because in UI we show this checkbox as disabled for non-Serverless, but looking at the settings object in the resource, we do have this field set to PERFORMANCE_OPTIMIZED

ilyakuz-db · 2026-03-07T12:36:21Z


-      # Same as clusters.node_type_id — see clusters/resource_cluster.go#L333
-      # s.SchemaPath("node_type_id").SetComputed()
-      - field: tasks[*].new_cluster.node_type_id


When I try to create new_cluster without node_type_id it fails with 400 error:

Endpoint: POST https://<host>/api/2.2/jobs/reset HTTP Status: 400 Bad Request API error_code: INVALID_PARAMETER_VALUE API message: Cluster validation error: Unknown node type id

Considering that backend default is applied only when config value is nil, and for nil value we have 400 error - it doesn't make sense to keep it here

Question, if I don't specify new_cluster, can it be added by the backend when resource is fetched? I guess not, or some test would fail, but good to say it explicitly.

In which case? You can either use existing_cluster_id fields or new_cluster object. node_type_id exists in new_cluster and it cannot be set independently. When you specify compute for a task, you have 3 options:

select existing cluster (existing_cluster_id field)

add new (new_cluster field)

use serverless (doesn't have node_type_id)

denik · 2026-03-09T13:21:10Z

    ignore_remote_changes:
+      # edit_mode is set by CLI, not user-configurable
+      - field: edit_mode
+        reason: managed


I'm using "managed" in other places to mean managed by backend, we should have another reason here.

denik · 2026-03-09T13:21:40Z


  jobs:
    ignore_remote_changes:
+      # edit_mode is set by CLI, not user-configurable


You mean DABs override user setting? Do you know if that is intentional?

No, I meant that the user can't specify this field in databricks.yml, it's not in the json schema

It's only set by CLI

Understood, although still not clear why ignore_remote_changes for it? What if CLI sets A and user updates to B, should we not act on that drift?

In config-remote-sync, we don't need to act, as this field only controls "Job is locked for edits" state. And when yaml sync feature is enabled, we always expect EDITABLE. We also set it explicitly during deployment:

cli/bundle/deploy/metadata/annotate_jobs.go

Line 35 in a16cdc1

job.EditMode = jobs.JobEditModeEditable

But for the bundle plan, perhaps it could make sense. Example use case - user clicks "Disconnect this job from DAB", which changes edit_mode: UI_LOCKED -> EDITABLE in the resource. When user deploys again - the change is edit_mode: EDITABLE -> UI_LOCKED

Maybe it makes sense to keep a separate option in resources.yml for such cases, for example: ignore_for_sync / ignore_in_sync

Moved to configsync package from here

denik · 2026-03-09T13:22:36Z

        reason: managed

    backend_defaults:
+      # Backend returns PERFORMANCE_OPTIMIZED for serverless jobs when not set


Is this field ever set for non-serverless jobs?

simonfaltum

Review (automated, 2 agents)

[Major] Migration drops suppression behavior for ~16 fields

bundle/direct/dresources/resources.yml

The old serverSideDefaults handled timeout_seconds, email_notifications, webhook_notifications, task disabled, pipeline_task.full_refresh, single_user_name, data_security_mode, pipeline continuous, and others. These have no equivalent rules in resources.yml yet. After this PR, config-remote-sync will start surfacing these as drift.

Additionally, removing node_type_id from backend_defaults changes deploy plan behavior (not just config-sync), potentially causing spurious updates for clusters where the backend auto-fills node type. Note the asymmetry: driver_node_type_id is still in backend_defaults.

Suggestion: Add equivalent backend_defaults/ignore_remote_changes rules for the missing fields, or explain why removing them is safe.

[Gap (Major)] No regression tests for migrated configsync classification rules

bundle/configsync/diff_test.go (deleted), bundle/configsync/diff.go

The only unit test file was deleted (90 lines of matchPattern tests). No replacement tests for shouldClassifySkip, filterEntityDefaults, or convertChangeDesc. The new glue logic (synthetic ChangeDesc{Old: nil, New: nil, Remote: value}) has subtle interactions with the classification pipeline.

Suggestion: Add table-driven tests for shouldClassifySkip and convertChangeDesc covering backend defaults, zero/nil values, nested maps, and non-default values.

[Suggestion] Document the hasConfigValue behavioral change

bundle/configsync/diff.go:107

The change from hasConfigValue := cd.Old != nil || cd.New != nil to hasConfigValue := cd.New != nil is a correctness fix (confirmed by the acceptance test update), but there's no comment explaining why only cd.New is checked, making it easy for a future contributor to "fix" it back.

[Nit] Silent error swallowing in shouldClassifySkip

bundle/configsync/diff.go:60-63 - Classification errors silently return false. A debug log would help diagnose issues.

[Nit] Array handling in filterEntityDefaults preserves nil entries

bundle/configsync/diff.go:72-77 - Filtered arrays produce [nil, nil, nil] instead of nil, inconsistent with the map case that returns nil for empty maps. Low priority.

## Changes Use the same logic for server-side defaults as in resources.yml logic. I have a PR that updates `config-remote-sync` to use `resources.yml`, but it's not yet ready to merge #4677 ## Why  Changes in performance mode were incorrectly ignored previously ## Tests

github-actions · 2026-05-18T01:10:36Z

This PR has not received an update in a while. If you want to keep this PR open, please leave a comment below or push a new commit and auto-close will be canceled.

## Changes Use the same logic for server-side defaults as in resources.yml logic. I have a PR that updates `config-remote-sync` to use `resources.yml`, but it's not yet ready to merge #4677 ## Why  Changes in performance mode were incorrectly ignored previously ## Tests

…adata config-remote-sync previously filtered server-side noise out of detected remote changes via a hand-maintained serverSideDefaults table in bundle/configsync/defaults.go. The table constantly lagged behind resources.yml. This change derives the filtering from the same lifecycle metadata that bundle plan uses (bundle/direct/dresources/resources.yml and resources.generated.yml) and deletes the table. configsync matches the metadata directly instead of relying on the plan's per-field actions: the plan answers "what will deploy do", which is not the same as "does this field belong in configuration". For example, ResourceDashboard.OverrideChangeDesc re-promotes etag drift to Update after the metadata chain (deploy needs that for modified-remotely detection), but etag must never be written to YAML. Concretely: - Fields matching ignore_remote_changes (in either the hand-written or the generated config) are never written to configuration, regardless of values or plan action. - Fields matching backend_defaults are dropped when absent from the config (cd.New == nil) and the remote value matches the rule, including the optional values constraint. - The same matching drives the nested filtering of entities added remotely as a whole (e.g. a new task); map fields that become empty after filtering are pruned, but empty elements inside arrays are kept as {} to preserve array arity. Two intentional semantic changes compared to the old table: - When a field set in config has a remote value that filters entirely to backend defaults, the operation is now Remove (the field is dropped from the YAML) instead of Replace with an empty map. This produces cleaner config and is pinned by a unit test. - Pre-existing backend_defaults entries without values constraints (e.g. notebook_task.source, data_security_mode) now make sync drop any remote value when the field is absent from config, where the old table only dropped one specific default value. This intentionally aligns sync with the deploy plan's backend-default semantics; adding values constraints to those entries would change plan behavior and is out of scope here. The matching helpers move from bundle/direct/bundle_plan.go into the dresources package (FindMatchingRule, MatchesAllowedValue, and the backend-defaults loop as ResourceLifecycleConfig.MatchesBackendDefault) so plan and sync share one implementation. Sync-specific rules stay in the configsync package, ported to structpath patterns keyed by resource type: jobs edit_mode (set by the CLI, not user-settable in databricks.yml), the jobs queue reset value, and name-prefix stripping. These have sync-only drift semantics and don't belong in resources.yml. One backend_defaults rule needs a sync-side exception rather than removal: the jobs new_cluster node_type_id entries (tasks[*] and job_clusters[*]) are computed fields for the plan. Config legitimately omits node_type_id when a cluster policy fixes it, and jobs/get then returns the policy-resolved value in the stored settings (verified on AWS prod, both with and without apply_policy_default_values), so removing the entries would cause permadrift on every plan and a spurious update on every deploy. For sync the opposite holds: a remotely added cluster's explicit node_type_id is required configuration, and stripping it via nested filtering would make the synced YAML undeployable (cluster creation fails with 400 "Unknown node type id" unless a pool or policy provides the node type). These fields are therefore listed in fieldsKeptForSync, which nested entity filtering consults before dropping a field. Keeping them unconditionally cannot introduce a node_type_id/instance_pool_id conflict: the Jobs API rejects requests carrying both, and jobs/get does not materialize node_type_id for pool-backed clusters (also verified empirically), so a remote cluster never has both fields. resources.yml changes: - jobs: add performance_target backend default with values ["PERFORMANCE_OPTIMIZED"]; the backend returns it for all jobs when unset, not only for serverless jobs. - jobs: add tasks[*].timeout_seconds / tasks[*].disabled backend defaults (and for_each_task variants) with values [0] / [false]. These are inert for the plan, which already skips them as all-empty; they are needed so config-remote-sync does not write these backend-returned defaults into YAML for tasks added remotely. - jobs: add email_notifications.no_alert_for_skipped_runs backend defaults with values [false] at the job level, tasks[*], and tasks[*].for_each_task.task. The backend populates this deprecated field when notifications are not configured. Inert for the plan: the all-empty check in addPerFieldActions runs before backend defaults and already skips zero values. Needed so sync drops the leaf and then prunes the empty email_notifications map for entities added remotely. - jobs: add tasks/job_clusters new_cluster.single_user_name backend defaults, mirroring the existing standalone clusters entry (#4418). - clusters: add driver_node_type_flexibility and worker_node_type_flexibility backend defaults; some backends compute alternate node types for standalone clusters when the config does not specify flexibility (observed on AWS staging workspaces). - dashboards: add etag to ignore_remote_changes (output_only). This does not change deploy behavior because OverrideChangeDesc runs after the metadata chain; see the comment in resources.yml. - pipelines: remove the clusters[*].node_type_id backend_defaults entry. Unlike jobs and standalone clusters, pipelines/get returns the stored spec verbatim: clusters relying on an instance pool or a policy-fixed node type never gain node_type_id in the spec (verified on AWS prod for both cases), and ResourcePipeline.DoRead reads the spec, so the entry can never match a real remote diff. Removing it also keeps sync's nested filtering from stripping node_type_id off remotely added pipeline clusters. - The jobs and standalone clusters node_type_id entries stay: jobs/get materializes policy-fixed node types (see above), and clusters/get reports the pool's node type at the top level for pool-backed clusters while config must omit it (the API rejects node_type_id together with instance_pool_id). Table entries that needed no resources.yml counterpart: zero/empty values (job and task timeout_seconds: 0 at the top level, empty email_notifications/webhook_notifications, pipelines continuous: false, tasks[*].pipeline_task.full_refresh: false) are already skipped by the plan's all-empty check before they reach sync, and empty maps inside remotely added entities are pruned by the nested filtering; the non-empty email_notifications form the backend returns ({no_alert_for_skipped_runs: false}) is covered by the new explicit rules above. The remaining entries (cloud attributes, data_security_mode, experiments artifact_location, registered_models owner/full_name/metastore_id/storage_location, volumes storage_location, sql_warehouses creator_name/min_num_clusters/warehouse_type, jobs run_as, pipelines storage) were already covered by existing resources.yml rules.

…er fields

…ster rules

ilyakuz-db temporarily deployed to test-trigger-is March 6, 2026 16:38 — with GitHub Actions Inactive

ilyakuz-db temporarily deployed to test-trigger-is March 7, 2026 11:01 — with GitHub Actions Inactive

ilyakuz-db commented Mar 7, 2026

View reviewed changes

ilyakuz-db temporarily deployed to test-trigger-is March 8, 2026 23:18 — with GitHub Actions Inactive

ilyakuz-db changed the title ~~Config sync migrate to resource yml~~ Use resource.yml in field classification in config-remote-sync Mar 9, 2026

ilyakuz-db marked this pull request as ready for review March 9, 2026 09:54

ilyakuz-db requested review from andrewnester, anton-107, denik, pietern, shreyas-goenka and simonfaltum as code owners March 9, 2026 09:54

denik reviewed Mar 9, 2026

View reviewed changes

ilyakuz-db temporarily deployed to test-trigger-is March 10, 2026 13:53 — with GitHub Actions Inactive

ilyakuz-db temporarily deployed to test-trigger-is March 11, 2026 10:11 — with GitHub Actions Inactive

ilyakuz-db requested a review from denik March 11, 2026 11:03

simonfaltum reviewed Mar 12, 2026

View reviewed changes

denik approved these changes Mar 18, 2026

View reviewed changes

ilyakuz-db mentioned this pull request Apr 8, 2026

Fix config sync for perfomance mode and timeouts #4908

Merged

github-actions Bot added the Stale label May 18, 2026

ilyakuz-db mentioned this pull request Jun 16, 2026

configsync: don't sync policy-injected custom_tags and cluster_log_conf #5610

Merged

github-actions Bot closed this Jun 28, 2026

ilyakuz-db reopened this Jul 2, 2026

ilyakuz-db force-pushed the config-sync-migrate-to-resource-yml branch from 4045284 to 2d76234 Compare July 2, 2026 12:28

ilyakuz-db temporarily deployed to test-trigger-is July 2, 2026 12:29 — with GitHub Actions Inactive

ilyakuz-db added 2 commits July 2, 2026 16:07

Add acceptance test pinning plan suppression of policy-injected clust…

dc8ac14

…er fields

ilyakuz-db force-pushed the config-sync-migrate-to-resource-yml branch from 191100d to dc8ac14 Compare July 2, 2026 14:09

ilyakuz-db temporarily deployed to test-trigger-is July 2, 2026 14:10 — with GitHub Actions Inactive

Restore pipeline_task.full_refresh sync parity and pin standalone-clu…

f7b3de5

…ster rules

ilyakuz-db temporarily deployed to test-trigger-is July 2, 2026 14:13 — with GitHub Actions Inactive

github-actions Bot removed the Stale label Jul 3, 2026

Uh oh!

Conversation

ilyakuz-db commented Mar 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changes

Why

Tests

Uh oh!

eng-dev-ecosystem-bot commented Mar 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Integration test report

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

simonfaltum left a comment

Choose a reason for hiding this comment

Review (automated, 2 agents)

[Major] Migration drops suppression behavior for ~16 fields

[Gap (Major)] No regression tests for migrated configsync classification rules

[Suggestion] Document the hasConfigValue behavioral change

[Nit] Silent error swallowing in shouldClassifySkip

[Nit] Array handling in filterEntityDefaults preserves nil entries

Uh oh!

github-actions Bot commented May 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

ilyakuz-db commented Mar 6, 2026 •

edited

Loading

eng-dev-ecosystem-bot commented Mar 6, 2026 •

edited

Loading