1. Proposed title of this feature request

Allow configuration of persistent storage for file buffers for the Vector log collector

2. What is the nature and description of the request?

There is no solution available to configure persistent storage for the  cluster logging collector file buffers to avoid loss of data.

3. Why does the customer need this? (List the business requirements here)

By default, the fluent.conf uses directories under "/var/lib/fluentd" for storing file buffers pending transfer to long-term storage in ElasticSearch. While ElasticSearch can be configured with persistent storage, the collector file buffers can not, meaning that whenever collector pod restart, logs in buffer will be lost. if destination is down, collector pod might fill out memory then get restarted(OOM).

4. List any affected packages or components.

• OpenShift Cluster Logging
• fluentd/collector pods