Safer reconciler spec (#167)

* reconciler, spec: forbid concurrent executions Signed-off-by: Miguel Duarte Barroso <[email protected]> * reconciler, spec: remove complete pods We leave the default `failedJobsHistoryLimit` - 1 - to enable the user to debug the reconciler in case it fails. Signed-off-by: Miguel Duarte Barroso <[email protected]> * reconciler, spec: do not retry on error Since a new pod will be scheduled in 5 minutes, it is safer to just wait for the new reconcile cycle than insist on reconciling right now. This is important to prevent the cluster to clog further on constrained clusters. Signed-off-by: Miguel Duarte Barroso <[email protected]>
k8snetworkplumbingwg · Nov 29, 2021 · abd29e8 · abd29e8
1 parent 2b15ccc
commit abd29e8
Showing 1 changed file with 3 additions and 0 deletions.
diff --git a/doc/crds/ip-reconciler-job.yaml b/doc/crds/ip-reconciler-job.yaml
@@ -7,9 +7,12 @@ metadata:
     tier: node
     app: whereabouts
 spec:
+  concurrencyPolicy: Forbid
+  successfulJobsHistoryLimit: 0
   schedule: "*/5 * * * *"
   jobTemplate:
     spec:
+      backoffLimit: 0
       template:
         metadata:
           labels: