From e97d8e7bc2556c6d6a779c3a3219bbdd4fb39fe6 Mon Sep 17 00:00:00 2001 From: Rafia Sabih Date: Fri, 18 Sep 2026 13:01:37 +0200 Subject: [PATCH v1] Add pg_stat_subscription_stats.worker_launch_failure_count Track, per subscription, how many times logicalrep_worker_launch() failed because no free logical replication worker slot was available. This makes worker-slot exhaustion observable without a synchronous pre-check, and covers cases where no CREATE/ALTER SUBSCRIPTION is involved at all -- e.g. max_logical_replication_workers lowered before a restart, after which the launcher's own retry loop keeps failing to start already-enabled subscriptions. Being a cumulative counter, a nonzero value means the subscription has failed at least once since the last stats reset, not that it is currently affected; pair with pg_stat_subscription to tell the two apart. --- src/backend/catalog/system_views.sql | 1 + src/backend/replication/logical/launcher.c | 1 + src/backend/utils/activity/pgstat_subscription.c | 14 ++++++++++++++ src/backend/utils/adt/pgstatfuncs.c | 10 ++++++++-- src/include/catalog/catversion.h | 2 +- src/include/catalog/pg_proc.dat | 6 +++--- src/include/pgstat.h | 4 +++- src/test/regress/expected/rules.out | 3 ++- 8 files changed, 33 insertions(+), 8 deletions(-) diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql index ad340887f54..bb957087add 100644 --- a/src/backend/catalog/system_views.sql +++ b/src/backend/catalog/system_views.sql @@ -1550,6 +1550,7 @@ CREATE VIEW pg_stat_subscription_stats AS ss.apply_error_count, ss.sync_seq_error_count, ss.sync_table_error_count, + ss.worker_launch_failure_count, ss.confl_insert_exists, ss.confl_update_origin_differs, ss.confl_update_exists, diff --git a/src/backend/replication/logical/launcher.c b/src/backend/replication/logical/launcher.c index 5057a928adb..262d31054f9 100644 --- a/src/backend/replication/logical/launcher.c +++ b/src/backend/replication/logical/launcher.c @@ -464,6 +464,7 @@ retry: if (worker == NULL) { LWLockRelease(LogicalRepWorkerLock); + pgstat_report_subscription_worker_launch_failure(subid); ereport(WARNING, (errcode(ERRCODE_CONFIGURATION_LIMIT_EXCEEDED), errmsg("out of logical replication worker slots"), diff --git a/src/backend/utils/activity/pgstat_subscription.c b/src/backend/utils/activity/pgstat_subscription.c index 3eaf3e0390f..a6fd4f7edce 100644 --- a/src/backend/utils/activity/pgstat_subscription.c +++ b/src/backend/utils/activity/pgstat_subscription.c @@ -56,6 +56,20 @@ pgstat_report_subscription_error(Oid subid) } } +/* + * Report a worker error + */ +void +pgstat_report_subscription_worker_launch_failure(Oid subid) +{ + PgStat_EntryRef *entry_ref; + PgStat_BackendSubEntry *sub_entry; + + entry_ref = pgstat_prep_pending_entry(PGSTAT_KIND_SUBSCRIPTION, InvalidOid, subid, NULL); + sub_entry = (PgStat_BackendSubEntry *) entry_ref->pending; + sub_entry->worker_launch_failure_count++; +} + /* * Report a subscription conflict. */ diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c index 64b6f60516c..985103e25f4 100644 --- a/src/backend/utils/adt/pgstatfuncs.c +++ b/src/backend/utils/adt/pgstatfuncs.c @@ -2415,7 +2415,7 @@ pg_stat_get_replication_slot(PG_FUNCTION_ARGS) Datum pg_stat_get_subscription_stats(PG_FUNCTION_ARGS) { -#define PG_STAT_GET_SUBSCRIPTION_STATS_COLS 13 +#define PG_STAT_GET_SUBSCRIPTION_STATS_COLS 14 Oid subid = PG_GETARG_OID(0); TupleDesc tupdesc; Datum values[PG_STAT_GET_SUBSCRIPTION_STATS_COLS] = {0}; @@ -2455,6 +2455,8 @@ pg_stat_get_subscription_stats(PG_FUNCTION_ARGS) INT8OID, -1, 0); TupleDescInitEntry(tupdesc, (AttrNumber) 13, "stats_reset", TIMESTAMPTZOID, -1, 0); + TupleDescInitEntry(tupdesc, (AttrNumber) 14, "worker_launch_failure_count", + INT8OID, -1, 0); TupleDescFinalize(tupdesc); BlessTupleDesc(tupdesc); @@ -2486,8 +2488,12 @@ pg_stat_get_subscription_stats(PG_FUNCTION_ARGS) nulls[i] = true; else values[i] = TimestampTzGetDatum(subentry->stat_reset_timestamp); + i++; + + /* worker_launch_failure_count */ + values[i++] = Int64GetDatum(subentry->worker_launch_failure_count); - Assert(i + 1 == PG_STAT_GET_SUBSCRIPTION_STATS_COLS); + Assert(i == PG_STAT_GET_SUBSCRIPTION_STATS_COLS); /* Returns the record as Datum */ PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls))); diff --git a/src/include/catalog/catversion.h b/src/include/catalog/catversion.h index 6f3e526de96..6cb7dc2289b 100644 --- a/src/include/catalog/catversion.h +++ b/src/include/catalog/catversion.h @@ -57,6 +57,6 @@ */ /* yyyymmddN */ -#define CATALOG_VERSION_NO 202609152 +#define CATALOG_VERSION_NO 202609181 #endif diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat index f46427258e3..33512da3dfc 100644 --- a/src/include/catalog/pg_proc.dat +++ b/src/include/catalog/pg_proc.dat @@ -5742,9 +5742,9 @@ { oid => '6231', descr => 'statistics: information about subscription stats', proname => 'pg_stat_get_subscription_stats', provolatile => 's', proparallel => 'r', prorettype => 'record', proargtypes => 'oid', - proallargtypes => '{oid,oid,int8,int8,int8,int8,int8,int8,int8,int8,int8,int8,int8,timestamptz}', - proargmodes => '{i,o,o,o,o,o,o,o,o,o,o,o,o,o}', - proargnames => '{subid,subid,apply_error_count,sync_seq_error_count,sync_table_error_count,confl_insert_exists,confl_update_origin_differs,confl_update_exists,confl_update_deleted,confl_update_missing,confl_delete_origin_differs,confl_delete_missing,confl_multiple_unique_conflicts,stats_reset}', + proallargtypes => '{oid,oid,int8,int8,int8,int8,int8,int8,int8,int8,int8,int8,int8,timestamptz,int8}', + proargmodes => '{i,o,o,o,o,o,o,o,o,o,o,o,o,o,o}', + proargnames => '{subid,subid,apply_error_count,sync_seq_error_count,sync_table_error_count,confl_insert_exists,confl_update_origin_differs,confl_update_exists,confl_update_deleted,confl_update_missing,confl_delete_origin_differs,confl_delete_missing,confl_multiple_unique_conflicts,stats_reset,worker_launch_failure_count}', prosrc => 'pg_stat_get_subscription_stats' }, { oid => '6118', descr => 'statistics: information about subscription', proname => 'pg_stat_get_subscription', prorows => '10', proisstrict => 'f', diff --git a/src/include/pgstat.h b/src/include/pgstat.h index 187d82c96fe..411829a8a9b 100644 --- a/src/include/pgstat.h +++ b/src/include/pgstat.h @@ -116,6 +116,7 @@ typedef struct PgStat_BackendSubEntry PgStat_Counter sync_seq_error_count; PgStat_Counter sync_table_error_count; PgStat_Counter conflict_count[CONFLICT_NUM_TYPES]; + PgStat_Counter worker_launch_failure_count; } PgStat_BackendSubEntry; /* ---------- @@ -502,6 +503,7 @@ typedef struct PgStat_StatSubEntry PgStat_Counter sync_seq_error_count; PgStat_Counter sync_table_error_count; PgStat_Counter conflict_count[CONFLICT_NUM_TYPES]; + PgStat_Counter worker_launch_failure_count; TimestampTz stat_reset_timestamp; } PgStat_StatSubEntry; @@ -925,7 +927,7 @@ extern void pgstat_report_subscription_conflict(Oid subid, ConflictType type); extern void pgstat_create_subscription(Oid subid); extern void pgstat_drop_subscription(Oid subid); extern PgStat_StatSubEntry *pgstat_fetch_stat_subscription(Oid subid); - +extern void pgstat_report_subscription_worker_launch_failure(Oid subid); /* * Functions in pgstat_xact.c diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out index 4a8cc759d7b..03daa8521da 100644 --- a/src/test/regress/expected/rules.out +++ b/src/test/regress/expected/rules.out @@ -2316,6 +2316,7 @@ pg_stat_subscription_stats| SELECT ss.subid, ss.apply_error_count, ss.sync_seq_error_count, ss.sync_table_error_count, + ss.worker_launch_failure_count, ss.confl_insert_exists, ss.confl_update_origin_differs, ss.confl_update_exists, @@ -2326,7 +2327,7 @@ pg_stat_subscription_stats| SELECT ss.subid, ss.confl_multiple_unique_conflicts, ss.stats_reset FROM pg_subscription s, - LATERAL pg_stat_get_subscription_stats(s.oid) ss(subid, apply_error_count, sync_seq_error_count, sync_table_error_count, confl_insert_exists, confl_update_origin_differs, confl_update_exists, confl_update_deleted, confl_update_missing, confl_delete_origin_differs, confl_delete_missing, confl_multiple_unique_conflicts, stats_reset); + LATERAL pg_stat_get_subscription_stats(s.oid) ss(subid, apply_error_count, sync_seq_error_count, sync_table_error_count, confl_insert_exists, confl_update_origin_differs, confl_update_exists, confl_update_deleted, confl_update_missing, confl_delete_origin_differs, confl_delete_missing, confl_multiple_unique_conflicts, stats_reset, worker_launch_failure_count); pg_stat_sys_indexes| SELECT relid, indexrelid, schemaname, -- 2.39.5 (Apple Git-154)