ceph-mixin: fix linting issue and add cluster template support

Fix most of the issues reported by dashboards-linter: - Add matcher/template for job (and also cluster) - use $__rate_interval everywhere Also this change all the irate functions to rate as most of irate where not actually used correctly. While using irate on graph for instance you can easily miss some of the metrics values as irate only take the two last values and the query steps can be quite large if you want a graph for a few hours/a day or more. Fixes: https://tracker.ceph.com/issues/55003 Signed-off-by: Arthur Outhenin-Chalandre <[email protected]> ceph-mixin: add config with matchers and tags Signed-off-by: Arthur Outhenin-Chalandre <[email protected]>
selroc · May 16, 2022 · faeea8d · faeea8d
1 parent 1452311
commit faeea8d
Show file tree

Hide file tree

Showing 22 changed files with 1,341 additions and 600 deletions.
diff --git a/monitoring/ceph-mixin/config.libsonnet b/monitoring/ceph-mixin/config.libsonnet
@@ -1 +1,8 @@
-{}
+{
+  _config+:: {
+    dashboardTags: ['ceph-mixin'],
+
+    clusterLabel: 'cluster',
+    showMultiCluster: false,
+  },
+}
diff --git a/monitoring/ceph-mixin/dashboards/cephfs.libsonnet b/monitoring/ceph-mixin/dashboards/cephfs.libsonnet
@@ -1,5 +1,6 @@
 local g = import 'grafonnet/grafana.libsonnet';
 local u = import 'utils.libsonnet';
+local c = (import '../mixin.libsonnet')._config;
 
 {
   grafanaDashboards+:: {
@@ -28,7 +29,7 @@ local u = import 'utils.libsonnet';
         'now-1h',
         '15s',
         16,
-        [],
+        c.dashboardTags,
         '',
         {
           refresh_intervals: ['5s', '10s', '15s', '30s', '1m', '5m', '15m', '30m', '1h', '2h', '1d'],
@@ -55,10 +56,16 @@ local u = import 'utils.libsonnet';
       .addTemplate(
         g.template.datasource('datasource', 'prometheus', 'default', label='Data Source')
       )
+      .addTemplate(
+        u.addClusterTemplate()
+      )
+      .addTemplate(
+        u.addJobTemplate()
+      )
       .addTemplate(
         u.addTemplateSchema('mds_servers',
                             '$datasource',
-                            'label_values(ceph_mds_inodes, ceph_daemon)',
+                            'label_values(ceph_mds_inodes{%(matchers)s}, ceph_daemon)' % u.matchers(),
                             1,
                             true,
                             1,
@@ -71,15 +78,15 @@ local u = import 'utils.libsonnet';
           'MDS Workload - $mds_servers',
           'none',
           'Reads(-) / Writes (+)',
-          'sum(rate(ceph_objecter_op_r{ceph_daemon=~"($mds_servers).*"}[1m]))',
+          'sum(rate(ceph_objecter_op_r{%(matchers)s, ceph_daemon=~"($mds_servers).*"}[$__rate_interval]))' % u.matchers(),
           'Read Ops',
           0,
           1,
           12,
           9
         )
         .addTarget(u.addTargetSchema(
-          'sum(rate(ceph_objecter_op_w{ceph_daemon=~"($mds_servers).*"}[1m]))',
+          'sum(rate(ceph_objecter_op_w{%(matchers)s, ceph_daemon=~"($mds_servers).*"}[$__rate_interval]))' % u.matchers(),
           'Write Ops'
         ))
         .addSeriesOverride(
@@ -89,7 +96,7 @@ local u = import 'utils.libsonnet';
           'Client Request Load - $mds_servers',
           'none',
           'Client Requests',
-          'ceph_mds_server_handle_client_request{ceph_daemon=~"($mds_servers).*"}',
+          'ceph_mds_server_handle_client_request{%(matchers)s, ceph_daemon=~"($mds_servers).*"}' % u.matchers(),
           '{{ceph_daemon}}',
           12,
           1,

diff --git a/monitoring/ceph-mixin/dashboards/dashboards.libsonnet b/monitoring/ceph-mixin/dashboards/dashboards.libsonnet
@@ -1,3 +1,4 @@
+(import '../config.libsonnet') +
 (import 'cephfs.libsonnet') +
 (import 'host.libsonnet') +
 (import 'osd.libsonnet') +

diff --git a/monitoring/ceph-mixin/dashboards/host.libsonnet b/monitoring/ceph-mixin/dashboards/host.libsonnet