support sidecar containers feature (kubernetes 1.29)

[jackskj]

Crashing telegraf container (due to OOM or erroneous configuration) has effect on the main application as the state of the entire pod is set to "not ready" if telegraf is restarted.
In such occurrence, the entire pod will be taken out of kubernetes endpoints and service.

To overcome this issue, kubernetes 1.29 has a "sidecar container" beta feature.
With this feature, the lifecycle of the main pod is separate to that of any sicecar containers.

To implement this feature, telegraf-operator would have to be injected as an iniContiner with restartPolicy set to "always"

Relevant URLs

kubernetes 1.29 sidecar feature

What products and version are you using?

telegraf-operatror: v1.3.11
kubernetes v1.23 (planning to update)

[Pradyumnakashyap]

+1

I would like this feature to be added as well !
We use the Telegraf Operator to inject a side car into a Kubernetes Job and we see that the Telegraf container continues to run even after the main container completes.
The above feature would provide us with an option to match the lifecycle with the main container.

[grigory-bogush]

In our use case we have kubernetes jobs which never finish without timeout due to telegraf being one of the main containers, so it would be nice to have this feature

[0054]

+1

[D0Ge3]

+1

[chyvak357]

I can’t stop the Telegraf container without the whole pod going into "not ready" state. Tried everything, and all my searches led me here. Hoping a solution can be found :(

[dploeger]

Additionally, the mutating webhook removes "RestartPolicy: Always" from init containers when they are used by other apps.