The document says fluentd retries chunk flushing indefinitely.

Fluentd retries chunk flushing indefinitely. In OpenShift Container Platform, you cannot change the indefinite retry behavior.

But in OpenShift Logging 5.4.3, ConfigMap of a collector pod contains the following line, meaning the retry stops after 60 minutes.

retry_timeout 60m

I guess this behavior change has been introduced in https://issues.redhat.com/browse/LOG-1500.

The document needs to be updated accordingly.