diff --git a/.releasenotes/notes/add-prometheus-relay-df282e14ed88da8c.yaml b/.releasenotes/notes/add-prometheus-relay-df282e14ed88da8c.yaml
new file mode 100644
index 0000000..dbf8a39
--- /dev/null
+++ b/.releasenotes/notes/add-prometheus-relay-df282e14ed88da8c.yaml
@@ -0,0 +1,12 @@
+---
+
+summary: >
+  Add prometheus relay service
+
+features:
+  - The Prometheus Relay service is responsible for
+    getting PromQL queries from external components,
+    such as Grafana, passing them to all discovered
+    Prometheus servers, merging the results and returning
+    the data. Prometheus Relay can be used to handle
+    Prometheus high availability or sharding.
diff --git a/.releasenotes/notes/spawn-multiple-replicas-prometheus-b80eaede9c19b8cd.yaml b/.releasenotes/notes/spawn-multiple-replicas-prometheus-b80eaede9c19b8cd.yaml
new file mode 100644
index 0000000..cc11b73
--- /dev/null
+++ b/.releasenotes/notes/spawn-multiple-replicas-prometheus-b80eaede9c19b8cd.yaml
@@ -0,0 +1,14 @@
+summary: >
+  Spawned two replicas of Prometheus to provide HA
+
+features:
+  - Spawned two replicas of Prometheus inside Docker Swarm.
+    These two instances have the same configuration file
+    and scrape the same endpoints. Therefore, they can be
+    treated as one-to-one copies.
+    Due to a limitation, when connecting to the Prometheus
+    web UI it is not possible to choose to which of the
+    existing Prometheus servers to connect and those
+    servers may contain slightly different results for the
+    same queries. Therefore, we suggest that you use Grafana
+    to visualize the data.
diff --git a/docker/swarm/stack/monitoring/init.yml b/docker/swarm/stack/monitoring/init.yml
index 95f5f8d..e274426 100644
--- a/docker/swarm/stack/monitoring/init.yml
+++ b/docker/swarm/stack/monitoring/init.yml
@@ -15,6 +15,22 @@
               driver_opts:
                 encrypted: 1
           service:
+            relay:
+              networks:
+                - monitoring
+              deploy:
+                replicas: 2
+                labels:
+                  com.mirantis.monitoring: "relay"
+                restart_policy:
+                  condition: any
+              labels:
+                com.mirantis.monitoring: "relay"
+              image: ${_param:docker_image_prometheus_relay}
+              ports:
+                - 15016:8080
+              environment:
+                PROMETHEUS_RELAY_DNS: 'tasks.monitoring_server'
             remote_storage_adapter:
               networks:
                 - monitoring
@@ -81,7 +97,7 @@
               networks:
                 - monitoring
               deploy:
-                replicas: 1
+                replicas: 2
                 labels:
                   com.mirantis.monitoring: "prometheus"
                 restart_policy: