commit bda1c7bea2941640b1846a62a40aa4f01aaff24d
author Michal Kobus <mkobus@mirantis.com> Tue Sep 17 14:20:22 2019 +0200
committer mkobus <mkobus@mirantis.com> Mon Sep 23 10:48:42 2019 +0000
tree 140c1c7ecd592567871ddf009c1975c6dd17f6ca
parent e7a67d77bd7ae907640e2994f0e61ccdf987e96b

Add DockerdServiceReplicaFlapping alert

Change-Id: If388bc1091dc80f6b63f33e29384df177f55a937
Related-bug: PROD-33295
(cherry picked from commit 1a51036e93ebfd2bdab6c66aae32ffa955079308)

docker/meta/prometheus.yml

diff --git a/docker/meta/prometheus.yml b/docker/meta/prometheus.yml
index 6921039..0b1f345 100644
--- a/docker/meta/prometheus.yml
+++ b/docker/meta/prometheus.yml
@@ -12,6 +12,16 @@
       annotations:
         summary: "Dockerd process is down"
         description: "The dockerd process on the {{ $labels.host }} node is down."
+    DockerdServiceReplicaFlapping:
+      if: >-
+        sum(changes(docker_swarm_tasks_running[10m])) by (service_name) > 0
+      for: 15m
+      labels:
+        severity: critical
+        service: docker
+      annotations:
+        summary: "{{ $labels.service_name }} is flapping"
+        description: "Docker Swarm {{ $labels.service_name }} service replica is flapping for 15 minutes."
   {%- endraw %}
     DockerServiceOutage:
       if: >-