As per the linked customer ticket, the verbosity of the indexer logs make it difficult to debug issues on a live cluster. Logs are rolling over very quickly due to them flooded with scan errors which happen during indexer warmup as indexes may not be ready for scans.
We need the logs to maintain longer history (24 hours as a minimum) to make it viable to be able to perform root-cause analysis on issues.
The linked ticket contains the logs in question. Requesting we review the constitution of the indexer log file with a review to reducing the verbosity / frequency of certain messages.