Add cost-based vacuum delay time to progress views.

This commit adds the amount of time spent sleeping due to
cost-based delay to the pg_stat_progress_vacuum and
pg_stat_progress_analyze system views.  A new configuration
parameter named track_cost_delay_timing, which is off by default,
controls whether this information is gathered.  For vacuum, the
reported value includes the sleep time of any associated parallel
workers.  However, parallel workers only report their sleep time
once per second to avoid overloading the leader process.

Bumps catversion.

Author: Bertrand Drouvot <[email protected]>
Co-authored-by: Nathan Bossart <[email protected]>
Reviewed-by: Sami Imseih <[email protected]>
Reviewed-by: Robert Haas <[email protected]>
Reviewed-by: Masahiko Sawada <[email protected]>
Reviewed-by: Masahiro Ikeda <[email protected]>
Reviewed-by: Dilip Kumar <[email protected]>
Reviewed-by: Sergei Kornilov <[email protected]>
Discussion: https://postgr.es/m/ZmaXmWDL829fzAVX%40ip-10-97-1-34.eu-west-3.compute.internal

1 files changed, 64 insertions, 0 deletions

diff --git a/src/backend/commands/vacuum.c b/src/backend/commands/vacuum.c
index cddddb0f9ea..0239d9bae65 100644
--- a/src/backend/commands/vacuum.c
+++ b/src/backend/commands/vacuum.c
@@ -39,6 +39,7 @@
 #include "catalog/pg_inherits.h"
 #include "commands/cluster.h"
 #include "commands/defrem.h"
+#include "commands/progress.h"
 #include "commands/vacuum.h"
 #include "miscadmin.h"
 #include "nodes/makefuncs.h"
@@ -59,6 +60,12 @@
 #include "utils/snapmgr.h"
 #include "utils/syscache.h"
 
+/*
+ * Minimum interval for cost-based vacuum delay reports from a parallel worker.
+ * This aims to avoid sending too many messages and waking up the leader too
+ * frequently.
+ */
+#define PARALLEL_VACUUM_DELAY_REPORT_INTERVAL_NS	(NS_PER_S)
 
 /*
  * GUC parameters
@@ -70,6 +77,7 @@ int			vacuum_multixact_freeze_table_age;
 int			vacuum_failsafe_age;
 int			vacuum_multixact_failsafe_age;
 double		vacuum_max_eager_freeze_failure_rate;
+bool		track_cost_delay_timing;
 
 /*
  * Variables for cost-based vacuum delay. The defaults differ between
@@ -80,6 +88,9 @@ double		vacuum_max_eager_freeze_failure_rate;
 double		vacuum_cost_delay = 0;
 int			vacuum_cost_limit = 200;
 
+/* Variable for reporting cost-based vacuum delay from parallel workers. */
+int64		parallel_vacuum_worker_delay_ns = 0;
+
 /*
  * VacuumFailsafeActive is a defined as a global so that we can determine
  * whether or not to re-enable cost-based vacuum delay when vacuuming a table.
@@ -2416,13 +2427,66 @@ vacuum_delay_point(bool is_analyze)
 	/* Nap if appropriate */
 	if (msec > 0)
 	{
+		instr_time	delay_start;
+
 		if (msec > vacuum_cost_delay * 4)
 			msec = vacuum_cost_delay * 4;
 
+		if (track_cost_delay_timing)
+			INSTR_TIME_SET_CURRENT(delay_start);
+
 		pgstat_report_wait_start(WAIT_EVENT_VACUUM_DELAY);
 		pg_usleep(msec * 1000);
 		pgstat_report_wait_end();
 
+		if (track_cost_delay_timing)
+		{
+			instr_time	delay_end;
+			instr_time	delay;
+
+			INSTR_TIME_SET_CURRENT(delay_end);
+			INSTR_TIME_SET_ZERO(delay);
+			INSTR_TIME_ACCUM_DIFF(delay, delay_end, delay_start);
+
+			/*
+			 * For parallel workers, we only report the delay time every once
+			 * in a while to avoid overloading the leader with messages and
+			 * interrupts.
+			 */
+			if (IsParallelWorker())
+			{
+				static instr_time last_report_time;
+				instr_time	time_since_last_report;
+
+				Assert(!is_analyze);
+
+				/* Accumulate the delay time */
+				parallel_vacuum_worker_delay_ns += INSTR_TIME_GET_NANOSEC(delay);
+
+				/* Calculate interval since last report */
+				INSTR_TIME_SET_ZERO(time_since_last_report);
+				INSTR_TIME_ACCUM_DIFF(time_since_last_report, delay_end, last_report_time);
+
+				/* If we haven't reported in a while, do so now */
+				if (INSTR_TIME_GET_NANOSEC(time_since_last_report) >=
+					PARALLEL_VACUUM_DELAY_REPORT_INTERVAL_NS)
+				{
+					pgstat_progress_parallel_incr_param(PROGRESS_VACUUM_DELAY_TIME,
+														parallel_vacuum_worker_delay_ns);
+
+					/* Reset variables */
+					last_report_time = delay_end;
+					parallel_vacuum_worker_delay_ns = 0;
+				}
+			}
+			else if (is_analyze)
+				pgstat_progress_incr_param(PROGRESS_ANALYZE_DELAY_TIME,
+										   INSTR_TIME_GET_NANOSEC(delay));
+			else
+				pgstat_progress_incr_param(PROGRESS_VACUUM_DELAY_TIME,
+										   INSTR_TIME_GET_NANOSEC(delay));
+		}
+
 		/*
 		 * We don't want to ignore postmaster death during very long vacuums
 		 * with vacuum_cost_delay configured.  We can't use the usual