App health with CRD implementing cronjob

[mtotone]

Hello,

I try to manage mariadb-operator resources with an internal helm app and we're facing a technical issue with backup CRD.

Backup CRD is a cronjob and the app health is shown as progressing until cronjob is executed for the first time.

Healthy app:
NAMESPACE NAME COMPLETE STATUS MARIADB AGE <namespace> backup-scheduled True Success <cluster> 22h

Progressing app:
NAMESPACE NAME COMPLETE STATUS MARIADB AGE <namespace> backup-scheduled False Scheduled <cluster> 6m17s

For the progressing app, the cronjob last schedule is none.

For now, i use a lua script to manage mariadb-operator backup object and set status when reason is 'CronJobScheduled'.

Have you ever face a similar issue (with mariadb-operator or other) and how did you solve it ?

Thanks.

[HanssenKai]

Unless someone contributes a custom health check for mariaDB resources to the argoCD repo, then a lua override is the way to go. In the meantime, it would be nice if you could share the lua snippet for others who run in to the same issue.

By default, kubernetes cronJobs have empty status fields until they are run the first time, so a mariaDB resource health check would probably need an additional check for status.type == Complete and filter by status.reason == CronJobSuccess | CronJobFailed | CronJobScheduled