Unable to start metrics sidecar innodbcluster

[Thomas de Gier]

I recently upgraded the mysql-operator from 8.0.33 to 8.3.0 to use the added metrics feature for prometheus monitoring. I want to deploy the sidecars in my existing innodbclusters in kubernetes. The operator is updated and the existing innodbclusters seem to deploy with their previous manifest.

However, when applying the manifest with a metrics object, the metrics sidecar keeps crashing. I also tried this with a new test cluster. The applied manifest for this cluster:

apiVersion: mysql.oracle.com/v2
kind: InnoDBCluster
metadata:
name: ictest
namespace: testing
spec:
secretName: ictest-secret
tlsUseSelfSigned: true
datadirVolumeClaimTemplate:
accessModes:
- ReadWriteOnce
storageClassName: ceph-nvme-block
metrics:
enable: true
image: prom/mysqld-exporter
---
apiVersion: v1
kind: Secret
metadata:
name: ictest-secret
namespace: testing
stringData:
rootUser: mysql
rootHost: "%"
rootPassword: pass


Otherwise, the other containers seem functional and the cluster seems operational.

Is something wrong with this innodbcluster configuration? Can someone share their experience with correctly deploying the sidecar?