Alerts: migrate RequestErrors and RulerRemoteEvaluationFailing to nat…

…ive histograms (#9004) * Alerts: enrich RequestErrors and RulerRemoteEvaluationFailing with a native histogram version Signed-off-by: Yuri Nikolic <durica.nikolic@grafana.com> * Making lint happy Signed-off-by: Yuri Nikolic <durica.nikolic@grafana.com> * Update operations/mimir-mixin/alerts/alerts.libsonnet Co-authored-by: George Krajcsovits <krajorama@users.noreply.github.com> * Making lint happy Signed-off-by: Yuri Nikolic <durica.nikolic@grafana.com> --------- Signed-off-by: Yuri Nikolic <durica.nikolic@grafana.com> Co-authored-by: George Krajcsovits <krajorama@users.noreply.github.com>
grafana · Aug 15, 2024 · 3512a1d · 3512a1d
1 parent 1543093
commit 3512a1d
Show file tree

Hide file tree

Showing 6 changed files with 200 additions and 67 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -113,6 +113,7 @@
 * [ENHANCEMENT] Dashboards: remove "All" option for namespace dropdown in dashboards. #8829
 * [ENHANCEMENT] Dashboards: add Kafka end-to-end latency outliers panel in the "Mimir / Writes" dashboard. #8948
 * [ENHANCEMENT] Dashboards: add "Out-of-order samples appended" panel to "Mimir / Tenants" dashboard. #8939
+* [ENHANCEMENT] Alerts: `RequestErrors` and `RulerRemoteEvaluationFailing` have been enriched with a native histogram version. #9004
 * [BUGFIX] Dashboards: fix "current replicas" in autoscaling panels when HPA is not active. #8566
 * [BUGFIX] Alerts: do not fire `MimirRingMembersMismatch` during the migration to experimental ingest storage. #8727
 

diff --git a/operations/helm/charts/mimir-distributed/CHANGELOG.md b/operations/helm/charts/mimir-distributed/CHANGELOG.md
@@ -45,6 +45,7 @@ Entries should include a reference to the Pull Request that introduced the chang
 * [ENHANCEMENT] Add support for setting namespace for dashboard config maps. #8813
 * [ENHANCEMENT] Add support for string `extraObjects` for better support with templating. #8825
 * [ENHANCEMENT] Helm : allow setting a read and write urls to continous-test. #7674
+* [ENHANCEMENT] Alerts: `RequestErrors` and `RulerRemoteEvaluationFailing` have been enriched with a native histogram version. #9004
 * [BUGFIX] Add missing container security context to run `continuous-test` under the restricted security policy. #8653
 * [BUGFIX] Add `global.extraVolumeMounts` to the exporter container on memcached statefulsets #8787
 * [BUGFIX] Fix helm releases failing when `querier.kedaAutoscaling.predictiveScalingEnabled=true`. #8731

diff --git a/...amonitoring-values-generated/mimir-distributed/templates/metamonitoring/mixin-alerts.yaml b/...amonitoring-values-generated/mimir-distributed/templates/metamonitoring/mixin-alerts.yaml
@@ -32,12 +32,31 @@ spec:
               # - 529: used by distributor rate limiting (using 529 instead of 429 to let the client retry)
               # - 598: used by GEM gateway when the client is very slow to send the request and the gateway times out reading the request body
               (
-                sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{status_code=~"5..",status_code!~"529|598",route!~"ready|debug_pprof"}[1m]))
+                sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{status_code=~"5..", status_code!~"529|598", route!~"ready|debug_pprof"}[1m]))
                 /
                 sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{route!~"ready|debug_pprof"}[1m]))
               ) * 100 > 1
             for: 15m
             labels:
+              histogram: classic
+              severity: critical
+          - alert: MimirRequestErrors
+            annotations:
+              message: |
+                  The route {{ $labels.route }} in {{ $labels.cluster }}/{{ $labels.namespace }} is experiencing {{ printf "%.2f" $value }}% errors.
+              runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrequesterrors
+            expr: |
+              # The following 5xx errors considered as non-error:
+              # - 529: used by distributor rate limiting (using 529 instead of 429 to let the client retry)
+              # - 598: used by GEM gateway when the client is very slow to send the request and the gateway times out reading the request body
+              (
+                sum by (cluster, namespace, job, route) (histogram_count(rate(cortex_request_duration_seconds{status_code=~"5..", status_code!~"529|598", route!~"ready|debug_pprof"}[1m])))
+                /
+                sum by (cluster, namespace, job, route) (histogram_count(rate(cortex_request_duration_seconds{route!~"ready|debug_pprof"}[1m])))
+              ) * 100 > 1
+            for: 15m
+            labels:
+              histogram: native
               severity: critical
           - alert: MimirRequestLatency
             annotations:
@@ -485,13 +504,29 @@ spec:
                   Mimir rulers in {{ $labels.cluster }}/{{ $labels.namespace }} are failing to perform {{ printf "%.2f" $value }}% of remote evaluations through the ruler-query-frontend.
               runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrulerremoteevaluationfailing
             expr: |
-              100 * (
-              sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", status_code=~"5..", job=~".*/(ruler-query-frontend.*)"}[5m]))
+              (
+                sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{status_code=~"5..", route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
                 /
-              sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
-              ) > 1
+                sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
+              ) * 100 > 1
+            for: 5m
+            labels:
+              histogram: classic
+              severity: warning
+          - alert: MimirRulerRemoteEvaluationFailing
+            annotations:
+              message: |
+                  Mimir rulers in {{ $labels.cluster }}/{{ $labels.namespace }} are failing to perform {{ printf "%.2f" $value }}% of remote evaluations through the ruler-query-frontend.
+              runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrulerremoteevaluationfailing
+            expr: |
+              (
+                sum by (cluster, namespace) (histogram_count(rate(cortex_request_duration_seconds{status_code=~"5..", route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m])))
+                /
+                sum by (cluster, namespace) (histogram_count(rate(cortex_request_duration_seconds{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m])))
+              ) * 100 > 1
             for: 5m
             labels:
+              histogram: native
               severity: warning
       - name: gossip_alerts
         rules:

diff --git a/operations/mimir-mixin-compiled-baremetal/alerts.yaml b/operations/mimir-mixin-compiled-baremetal/alerts.yaml
@@ -20,12 +20,31 @@ groups:
             # - 529: used by distributor rate limiting (using 529 instead of 429 to let the client retry)
             # - 598: used by GEM gateway when the client is very slow to send the request and the gateway times out reading the request body
             (
-              sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{status_code=~"5..",status_code!~"529|598",route!~"ready|debug_pprof"}[1m]))
+              sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{status_code=~"5..", status_code!~"529|598", route!~"ready|debug_pprof"}[1m]))
               /
               sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{route!~"ready|debug_pprof"}[1m]))
             ) * 100 > 1
           for: 15m
           labels:
+            histogram: classic
+            severity: critical
+        - alert: MimirRequestErrors
+          annotations:
+            message: |
+                The route {{ $labels.route }} in {{ $labels.cluster }}/{{ $labels.namespace }} is experiencing {{ printf "%.2f" $value }}% errors.
+            runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrequesterrors
+          expr: |
+            # The following 5xx errors considered as non-error:
+            # - 529: used by distributor rate limiting (using 529 instead of 429 to let the client retry)
+            # - 598: used by GEM gateway when the client is very slow to send the request and the gateway times out reading the request body
+            (
+              sum by (cluster, namespace, job, route) (histogram_count(rate(cortex_request_duration_seconds{status_code=~"5..", status_code!~"529|598", route!~"ready|debug_pprof"}[1m])))
+              /
+              sum by (cluster, namespace, job, route) (histogram_count(rate(cortex_request_duration_seconds{route!~"ready|debug_pprof"}[1m])))
+            ) * 100 > 1
+          for: 15m
+          labels:
+            histogram: native
             severity: critical
         - alert: MimirRequestLatency
           annotations:
@@ -463,13 +482,29 @@ groups:
                 Mimir rulers in {{ $labels.cluster }}/{{ $labels.namespace }} are failing to perform {{ printf "%.2f" $value }}% of remote evaluations through the ruler-query-frontend.
             runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrulerremoteevaluationfailing
           expr: |
-            100 * (
-            sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", status_code=~"5..", job=~".*/(ruler-query-frontend.*)"}[5m]))
+            (
+              sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{status_code=~"5..", route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
               /
-            sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
-            ) > 1
+              sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
+            ) * 100 > 1
+          for: 5m
+          labels:
+            histogram: classic
+            severity: warning
+        - alert: MimirRulerRemoteEvaluationFailing
+          annotations:
+            message: |
+                Mimir rulers in {{ $labels.cluster }}/{{ $labels.namespace }} are failing to perform {{ printf "%.2f" $value }}% of remote evaluations through the ruler-query-frontend.
+            runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrulerremoteevaluationfailing
+          expr: |
+            (
+              sum by (cluster, namespace) (histogram_count(rate(cortex_request_duration_seconds{status_code=~"5..", route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m])))
+              /
+              sum by (cluster, namespace) (histogram_count(rate(cortex_request_duration_seconds{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m])))
+            ) * 100 > 1
           for: 5m
           labels:
+            histogram: native
             severity: warning
     - name: gossip_alerts
       rules:

diff --git a/operations/mimir-mixin-compiled/alerts.yaml b/operations/mimir-mixin-compiled/alerts.yaml
@@ -20,12 +20,31 @@ groups:
             # - 529: used by distributor rate limiting (using 529 instead of 429 to let the client retry)
             # - 598: used by GEM gateway when the client is very slow to send the request and the gateway times out reading the request body
             (
-              sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{status_code=~"5..",status_code!~"529|598",route!~"ready|debug_pprof"}[1m]))
+              sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{status_code=~"5..", status_code!~"529|598", route!~"ready|debug_pprof"}[1m]))
               /
               sum by (cluster, namespace, job, route) (rate(cortex_request_duration_seconds_count{route!~"ready|debug_pprof"}[1m]))
             ) * 100 > 1
           for: 15m
           labels:
+            histogram: classic
+            severity: critical
+        - alert: MimirRequestErrors
+          annotations:
+            message: |
+                The route {{ $labels.route }} in {{ $labels.cluster }}/{{ $labels.namespace }} is experiencing {{ printf "%.2f" $value }}% errors.
+            runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrequesterrors
+          expr: |
+            # The following 5xx errors considered as non-error:
+            # - 529: used by distributor rate limiting (using 529 instead of 429 to let the client retry)
+            # - 598: used by GEM gateway when the client is very slow to send the request and the gateway times out reading the request body
+            (
+              sum by (cluster, namespace, job, route) (histogram_count(rate(cortex_request_duration_seconds{status_code=~"5..", status_code!~"529|598", route!~"ready|debug_pprof"}[1m])))
+              /
+              sum by (cluster, namespace, job, route) (histogram_count(rate(cortex_request_duration_seconds{route!~"ready|debug_pprof"}[1m])))
+            ) * 100 > 1
+          for: 15m
+          labels:
+            histogram: native
             severity: critical
         - alert: MimirRequestLatency
           annotations:
@@ -473,13 +492,29 @@ groups:
                 Mimir rulers in {{ $labels.cluster }}/{{ $labels.namespace }} are failing to perform {{ printf "%.2f" $value }}% of remote evaluations through the ruler-query-frontend.
             runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrulerremoteevaluationfailing
           expr: |
-            100 * (
-            sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", status_code=~"5..", job=~".*/(ruler-query-frontend.*)"}[5m]))
+            (
+              sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{status_code=~"5..", route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
               /
-            sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
-            ) > 1
+              sum by (cluster, namespace) (rate(cortex_request_duration_seconds_count{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m]))
+            ) * 100 > 1
+          for: 5m
+          labels:
+            histogram: classic
+            severity: warning
+        - alert: MimirRulerRemoteEvaluationFailing
+          annotations:
+            message: |
+                Mimir rulers in {{ $labels.cluster }}/{{ $labels.namespace }} are failing to perform {{ printf "%.2f" $value }}% of remote evaluations through the ruler-query-frontend.
+            runbook_url: https://grafana.com/docs/mimir/latest/operators-guide/mimir-runbooks/#mimirrulerremoteevaluationfailing
+          expr: |
+            (
+              sum by (cluster, namespace) (histogram_count(rate(cortex_request_duration_seconds{status_code=~"5..", route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m])))
+              /
+              sum by (cluster, namespace) (histogram_count(rate(cortex_request_duration_seconds{route="/httpgrpc.HTTP/Handle", job=~".*/(ruler-query-frontend.*)"}[5m])))
+            ) * 100 > 1
           for: 5m
           labels:
+            histogram: native
             severity: warning
     - name: gossip_alerts
       rules: