Containerd log format does not work properyl with current fluentd config

[pipo02mix]

Is your feature request related to a problem? Please describe.

Since we have switched from docker-shim to containerd the log format has changed and now the logs are processed by EFK weirdly in fluentd logging app.

Describe the solution you'd like

We need to change fluentd collector to parse the new format correctly
fluent/fluentd-kubernetes-daemonset#412

Describe alternatives you've considered

Additional context

[weatherhog]

@QuentinBisson do you know how to solve this issue?

[QuentinBisson]

This requires us to upgrade the fluent-bit app in the fluent logshipping app and to add the new cri parser that exists upstream and configure it for the specific kubernetes versions

[pipo02mix]

This worked for or customer FYI

        @type cri
        <parse>
          @type json
          time_key time
          time_format %Y-%m-%dT%H:%M:%S.%NZ
          keep_time_key true
        </parse>
       </parse>```

[QuentinBisson]

@giantswarm/team-phoenix do we have a way to know if we are usign docker or CRI? Either a kubernetes version, or smth else? Otherwise we will have to enforce customer configuration

[whites11]

You have it in the node CR info.

[QuentinBisson]

Thanks I'll see if I can access that using helm .Capabilities, otherwise, can I assume versions 18+ are using containerd?

[whites11]

yes

[QuentinBisson]

Fixed by giantswarm/fluent-logshipping-app#93