diff --git a/docs/production/deployment.md b/docs/production/deployment.md
index d710a628d5..846feed0bb 100644
--- a/docs/production/deployment.md
+++ b/docs/production/deployment.md
@@ -492,6 +492,20 @@ non-empty value is currently equivalent to true).
 
 [s3-uploads]: ../production/upload-backends.html#s3-backend-configuration
 
+#### `queue_workers_multiprocess`
+
+By default, Zulip automatically detects whether the system has enough
+memory to run Zulip queue processors in the higher-throughput but more
+multiprocess mode (or to save 1.5GiB of RAM with the multithreaded
+mode). The calculation is based on whether the system has enough
+memory (currently 3.5GiB) to run a single-server Zulip installation in
+the multiprocess mode.
+
+Set to `true` or `false` to override the automatic calculation.  This
+override is useful both Docker systems (where the above algorithm
+might see the host's memory, not the container's) and/or when using
+remote servers for postgres, memcached, redis, and RabbitMQ.
+
 #### `uwsgi_buffer_size`
 
 Override the default uwsgi buffer size of 8192.
diff --git a/puppet/zulip/manifests/app_frontend_base.pp b/puppet/zulip/manifests/app_frontend_base.pp
index def8949f7f..86fc22d2ab 100644
--- a/puppet/zulip/manifests/app_frontend_base.pp
+++ b/puppet/zulip/manifests/app_frontend_base.pp
@@ -72,7 +72,8 @@ class zulip::app_frontend_base {
   # This determines whether we run queue processors multithreaded or
   # multiprocess.  Multiprocess scales much better, but requires more
   # RAM; we just auto-detect based on available system RAM.
-  $queues_multiprocess = $zulip::common::total_memory_mb > 3500
+  $queues_multiprocess_default = $zulip::common::total_memory_mb > 3500
+  $queues_multiprocess = Boolean(zulipconf('application_server', 'queue_workers_multiprocess', $queues_multiprocess_default))
   $queues = [
     'deferred_work',
     'digest_emails',