add job restarts limit to values (#54)

Nextdoor · Sep 16, 2021 · c378f19 · c378f19
1 parent 8359558
commit c378f19
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 6 deletions.
diff --git a/charts/flink-job/Chart.yaml b/charts/flink-job/Chart.yaml
@@ -2,7 +2,7 @@ apiVersion: v2
 appVersion: "1.0"
 description: Flink job cluster on k8s
 name: flink-job
-version: 0.0.4
+version: 0.0.5
 maintainers:
   - name: Zedive
     email: albert@nextdoor.com
diff --git a/charts/flink-job/README.md b/charts/flink-job/README.md
@@ -2,7 +2,7 @@
 
 Flink job cluster on k8s
 
-![Version: 0.0.4](https://img.shields.io/badge/Version-0.0.4-informational?style=flat-square) ![AppVersion: 1.0](https://img.shields.io/badge/AppVersion-1.0-informational?style=flat-square)
+![Version: 0.0.5](https://img.shields.io/badge/Version-0.0.5-informational?style=flat-square) ![AppVersion: 1.0](https://img.shields.io/badge/AppVersion-1.0-informational?style=flat-square)
 
 This chart deploys a flink job cluster and runs a simple word counting flink app as an example.
 This chart includes some production ready set-ups such as
@@ -19,8 +19,9 @@ See metrics reporter in the flink properties for more details.
 
 | Key | Type | Default | Description |
 |-----|------|---------|-------------|
-| alerts.enabled | bool | `true` | (Boolean) whether to create the PrometheusRule for this flink cluster |
-| alerts.severity | string | `"info"` |  |
+| alerts.enabled | bool | `true` | (Boolean) Specifies whether to create the PrometheusRule for this flink cluster |
+| alerts.restartsLimit | int | `2` | (`int`) The number of job restarts before alerting |
+| alerts.severity | string | `"info"` | (String) Severity of the alerts |
 | defaults.runbookUrl | string | `"https://github.com/Nextdoor/k8s-charts/blob/main/charts/flink-job/runbook.md"` | (String) Runbook URL for the Prometheus alerts |
 | envVars | list | `[{"name":"HADOOP_CLASSPATH","value":"/opt/flink/opt/flink-metrics-prometheus-1.9.3.jar"}]` | Environment variables shared by all containers |
 | flinkProperties | object | `{"execution.checkpointing.interval":"10min","execution.checkpointing.mode":"EXACTLY_ONCE","high-availability":"org.apache.flink.kubernetes.highavailability.KubernetesHaServicesFactory","high-availability.storageDir":"file:/savepoint/","kubernetes.cluster-id":"{{ .Values.fullnameOverride }}","kubernetes.namespace":"{{ .Release.Namespace }}","metrics.reporter.prom.class":"org.apache.flink.metrics.prometheus.PrometheusReporter","metrics.reporters":"prom","restart-strategy":"exponential-delay","restart-strategy.exponential-delay.backoff-multiplier":"2.0","state.checkpoints.dir":"file:/savepoint/","taskmanager.numberOfTaskSlots":"1"}` | (`Map`) Flink properties which are appened to flink-conf.yaml |

diff --git a/charts/flink-job/templates/prometheusrule.yaml b/charts/flink-job/templates/prometheusrule.yaml
@@ -47,7 +47,7 @@ spec:
         changes(flink_jobmanager_job_numRestarts{
           cluster="{{ $cluster }}",
           namespace="{{ $namespace }}"
-        }[30m]) > 2
+        }[30m]) > {{ .Values.alerts.restartsLimit }}
       for: 10m
       labels:
         severity: {{ .Values.alerts.severity }}

diff --git a/charts/flink-job/values.yaml b/charts/flink-job/values.yaml
@@ -199,9 +199,12 @@ savepoints:
   enabled: true
 
 alerts:
-  # -- (Boolean) whether to create the PrometheusRule for this flink cluster
+  # -- (Boolean) Specifies whether to create the PrometheusRule for this flink cluster
   enabled: true
+  # -- (String) Severity of the alerts
   severity: info
+  # -- (`int`) The number of job restarts before alerting
+  restartsLimit: 2
 
 defaults:
   # -- (String) Runbook URL for the Prometheus alerts