diff --git a/include/jemalloc/internal/background_thread_structs.h b/include/jemalloc/internal/background_thread_structs.h
index c02aa434..249115c3 100644
--- a/include/jemalloc/internal/background_thread_structs.h
+++ b/include/jemalloc/internal/background_thread_structs.h
@@ -48,6 +48,7 @@ struct background_thread_stats_s {
 	size_t num_threads;
 	uint64_t num_runs;
 	nstime_t run_interval;
+	mutex_prof_data_t max_counter_per_bg_thd;
 };
 typedef struct background_thread_stats_s background_thread_stats_t;
 
diff --git a/include/jemalloc/internal/mutex.h b/include/jemalloc/internal/mutex.h
index 7c24f072..f5b1163a 100644
--- a/include/jemalloc/internal/mutex.h
+++ b/include/jemalloc/internal/mutex.h
@@ -245,22 +245,25 @@ malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex) {
 	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
 }
 
-/* Copy the prof data from mutex for processing. */
 static inline void
-malloc_mutex_prof_read(tsdn_t *tsdn, mutex_prof_data_t *data,
-    malloc_mutex_t *mutex) {
-	mutex_prof_data_t *source = &mutex->prof_data;
-	/* Can only read holding the mutex. */
-	malloc_mutex_assert_owner(tsdn, mutex);
-
+malloc_mutex_prof_copy(mutex_prof_data_t *dst, mutex_prof_data_t *source) {
 	/*
 	 * Not *really* allowed (we shouldn't be doing non-atomic loads of
 	 * atomic data), but the mutex protection makes this safe, and writing
 	 * a member-for-member copy is tedious for this situation.
 	 */
-	*data = *source;
+	*dst = *source;
 	/* n_wait_thds is not reported (modified w/o locking). */
-	atomic_store_u32(&data->n_waiting_thds, 0, ATOMIC_RELAXED);
+	atomic_store_u32(&dst->n_waiting_thds, 0, ATOMIC_RELAXED);
+}
+
+/* Copy the prof data from mutex for processing. */
+static inline void
+malloc_mutex_prof_read(tsdn_t *tsdn, mutex_prof_data_t *data,
+    malloc_mutex_t *mutex) {
+	/* Can only read holding the mutex. */
+	malloc_mutex_assert_owner(tsdn, mutex);
+	malloc_mutex_prof_copy(data, &mutex->prof_data);
 }
 
 static inline void
@@ -285,4 +288,36 @@ malloc_mutex_prof_accum(tsdn_t *tsdn, mutex_prof_data_t *data,
 	data->n_lock_ops += source->n_lock_ops;
 }
 
+/* Compare the prof data and update to the maximum. */
+static inline void
+malloc_mutex_prof_max_update(tsdn_t *tsdn, mutex_prof_data_t *data,
+    malloc_mutex_t *mutex) {
+	mutex_prof_data_t *source = &mutex->prof_data;
+	/* Can only read holding the mutex. */
+	malloc_mutex_assert_owner(tsdn, mutex);
+
+	if (nstime_compare(&source->tot_wait_time, &data->tot_wait_time) > 0) {
+		nstime_copy(&data->tot_wait_time, &source->tot_wait_time);
+	}
+	if (nstime_compare(&source->max_wait_time, &data->max_wait_time) > 0) {
+		nstime_copy(&data->max_wait_time, &source->max_wait_time);
+	}
+	if (source->n_wait_times > data->n_wait_times) {
+		data->n_wait_times = source->n_wait_times;
+	}
+	if (source->n_spin_acquired > data->n_spin_acquired) {
+		data->n_spin_acquired = source->n_spin_acquired;
+	}
+	if (source->max_n_thds > data->max_n_thds) {
+		data->max_n_thds = source->max_n_thds;
+	}
+	if (source->n_owner_switches > data->n_owner_switches) {
+		data->n_owner_switches = source->n_owner_switches;
+	}
+	if (source->n_lock_ops > data->n_lock_ops) {
+		data->n_lock_ops = source->n_lock_ops;
+	}
+	/* n_wait_thds is not reported. */
+}
+
 #endif /* JEMALLOC_INTERNAL_MUTEX_H */
diff --git a/include/jemalloc/internal/mutex_prof.h b/include/jemalloc/internal/mutex_prof.h
index 6288ede5..190402e6 100644
--- a/include/jemalloc/internal/mutex_prof.h
+++ b/include/jemalloc/internal/mutex_prof.h
@@ -7,6 +7,7 @@
 
 #define MUTEX_PROF_GLOBAL_MUTEXES					\
     OP(background_thread)						\
+    OP(max_per_bg_thd)							\
     OP(ctl)								\
     OP(prof)								\
     OP(prof_thds_data)							\
diff --git a/src/background_thread.c b/src/background_thread.c
index 57b9b256..bea445f2 100644
--- a/src/background_thread.c
+++ b/src/background_thread.c
@@ -794,9 +794,11 @@ background_thread_stats_read(tsdn_t *tsdn, background_thread_stats_t *stats) {
 		return true;
 	}
 
-	stats->num_threads = n_background_threads;
-	uint64_t num_runs = 0;
 	nstime_init(&stats->run_interval, 0);
+	memset(&stats->max_counter_per_bg_thd, 0, sizeof(mutex_prof_data_t));
+
+	uint64_t num_runs = 0;
+	stats->num_threads = n_background_threads;
 	for (unsigned i = 0; i < max_background_threads; i++) {
 		background_thread_info_t *info = &background_thread_info[i];
 		if (malloc_mutex_trylock(tsdn, &info->mtx)) {
@@ -809,6 +811,8 @@ background_thread_stats_read(tsdn_t *tsdn, background_thread_stats_t *stats) {
 		if (info->state != background_thread_stopped) {
 			num_runs += info->tot_n_runs;
 			nstime_add(&stats->run_interval, &info->tot_sleep_time);
+			malloc_mutex_prof_max_update(tsdn,
+			    &stats->max_counter_per_bg_thd, &info->mtx);
 		}
 		malloc_mutex_unlock(tsdn, &info->mtx);
 	}
diff --git a/src/ctl.c b/src/ctl.c
index 0beef6e0..3ec6ca24 100644
--- a/src/ctl.c
+++ b/src/ctl.c
@@ -1042,6 +1042,9 @@ ctl_background_thread_stats_read(tsdn_t *tsdn) {
 		memset(stats, 0, sizeof(background_thread_stats_t));
 		nstime_init(&stats->run_interval, 0);
 	}
+	malloc_mutex_prof_copy(
+	    &ctl_stats->mutex_prof_data[global_prof_mutex_max_per_bg_thd],
+	    &stats->max_counter_per_bg_thd);
 }
 
 static void