diff --git a/src/backend/access/heap/vacuumlazy.c b/src/backend/access/heap/vacuumlazy.c
index 03c43efc32..3a188574aa 100644
--- a/src/backend/access/heap/vacuumlazy.c
+++ b/src/backend/access/heap/vacuumlazy.c
@@ -65,6 +65,7 @@
 #include "commands/dbcommands.h"
 #include "commands/progress.h"
 #include "commands/vacuum.h"
+#include "executor/instrument.h"
 #include "miscadmin.h"
 #include "optimizer/paths.h"
 #include "pgstat.h"
@@ -137,6 +138,7 @@
 #define PARALLEL_VACUUM_KEY_SHARED			1
 #define PARALLEL_VACUUM_KEY_DEAD_TUPLES		2
 #define PARALLEL_VACUUM_KEY_QUERY_TEXT		3
+#define PARALLEL_VACUUM_KEY_BUFFER_USAGE	4
 
 /*
  * Macro to check if we are in a parallel vacuum.  If true, we are in the
@@ -259,6 +261,9 @@ typedef struct LVParallelState
 	/* Shared information among parallel vacuum workers */
 	LVShared   *lvshared;
 
+	/* Points to buffer usage area in DSM */
+	BufferUsage	*buffer_usage;
+
 	/*
 	 * The number of indexes that support parallel index bulk-deletion and
 	 * parallel index cleanup respectively.
@@ -1991,6 +1996,7 @@ lazy_parallel_vacuum_indexes(Relation *Irel, IndexBulkDeleteResult **stats,
 							 int nindexes)
 {
 	int			nworkers;
+	int			i;
 
 	Assert(!IsParallelWorker());
 	Assert(ParallelVacuumIsActive(lps));
@@ -2085,8 +2091,18 @@ lazy_parallel_vacuum_indexes(Relation *Irel, IndexBulkDeleteResult **stats,
 	parallel_vacuum_index(Irel, stats, lps->lvshared,
 						  vacrelstats->dead_tuples, nindexes);
 
-	/* Wait for all vacuum workers to finish */
-	WaitForParallelWorkersToFinish(lps->pcxt);
+	/*
+	 * Next, accumulate buffer usage.  (This must wait for the workers to
+	 * finish, or we might get incomplete data.)
+	 */
+	if (nworkers > 0)
+	{
+		/* Wait for all vacuum workers to finish */
+		WaitForParallelWorkersToFinish(lps->pcxt);
+
+		for (i = 0; i < lps->pcxt->nworkers_launched; i++)
+			InstrAccumParallelQuery(&lps->buffer_usage[i]);
+	}
 
 	/*
 	 * Carry the shared balance value to heap scan and disable shared costing
@@ -3071,6 +3087,7 @@ begin_parallel_vacuum(Oid relid, Relation *Irel, LVRelStats *vacrelstats,
 	ParallelContext *pcxt;
 	LVShared   *shared;
 	LVDeadTuples *dead_tuples;
+	BufferUsage *buffer_usage;
 	bool	   *can_parallel_vacuum;
 	long		maxtuples;
 	char	   *sharedquery;
@@ -3154,6 +3171,18 @@ begin_parallel_vacuum(Oid relid, Relation *Irel, LVRelStats *vacrelstats,
 	shm_toc_estimate_chunk(&pcxt->estimator, est_deadtuples);
 	shm_toc_estimate_keys(&pcxt->estimator, 1);
 
+	/*
+	 * Estimate space for BufferUsage -- PARALLEL_VACUUM_KEY_BUFFER_USAGE.
+	 *
+	 * BufferUsage during executing maintenance command can be used by an
+	 * extension that reports the buffer usage, such as pg_stat_statements.
+	 * We have no way of knowing whether anyone's looking at pgBufferUsage,
+	 * so do it unconditionally.
+	 */
+	shm_toc_estimate_chunk(&pcxt->estimator,
+						   mul_size(sizeof(BufferUsage), pcxt->nworkers));
+	shm_toc_estimate_keys(&pcxt->estimator, 1);
+
 	/* Finally, estimate PARALLEL_VACUUM_KEY_QUERY_TEXT space */
 	querylen = strlen(debug_query_string);
 	shm_toc_estimate_chunk(&pcxt->estimator, querylen + 1);
@@ -3188,6 +3217,12 @@ begin_parallel_vacuum(Oid relid, Relation *Irel, LVRelStats *vacrelstats,
 	shm_toc_insert(pcxt->toc, PARALLEL_VACUUM_KEY_DEAD_TUPLES, dead_tuples);
 	vacrelstats->dead_tuples = dead_tuples;
 
+	/* Allocate space for each worker's BufferUsage; no need to initialize */
+	buffer_usage = shm_toc_allocate(pcxt->toc,
+									mul_size(sizeof(BufferUsage), pcxt->nworkers));
+	shm_toc_insert(pcxt->toc, PARALLEL_VACUUM_KEY_BUFFER_USAGE, buffer_usage);
+	lps->buffer_usage = buffer_usage;
+
 	/* Store query string for workers */
 	sharedquery = (char *) shm_toc_allocate(pcxt->toc, querylen + 1);
 	memcpy(sharedquery, debug_query_string, querylen + 1);
@@ -3317,6 +3352,7 @@ parallel_vacuum_main(dsm_segment *seg, shm_toc *toc)
 	Relation   *indrels;
 	LVShared   *lvshared;
 	LVDeadTuples *dead_tuples;
+	BufferUsage	*buffer_usage;
 	int			nindexes;
 	char	   *sharedquery;
 	IndexBulkDeleteResult **stats;
@@ -3369,9 +3405,16 @@ parallel_vacuum_main(dsm_segment *seg, shm_toc *toc)
 	if (lvshared->maintenance_work_mem_worker > 0)
 		maintenance_work_mem = lvshared->maintenance_work_mem_worker;
 
+	/* Prepare to track buffer usage during parallel execution */
+	InstrStartParallelQuery();
+
 	/* Process indexes to perform vacuum/cleanup */
 	parallel_vacuum_index(indrels, stats, lvshared, dead_tuples, nindexes);
 
+	/* Report buffer usage during parallel execution */
+	buffer_usage = shm_toc_lookup(toc, PARALLEL_VACUUM_KEY_BUFFER_USAGE, false);
+	InstrEndParallelQuery(&buffer_usage[ParallelWorkerNumber]);
+
 	vac_close_indexes(nindexes, indrels, RowExclusiveLock);
 	table_close(onerel, ShareUpdateExclusiveLock);
 	pfree(stats);