Agent fails startup during logging of messages exceeding the max-log-size parameterDescriptionLogging on the agent-client-collector has the potential to go into a continuous loop of creating and archiving zero length log files when auto-mid-selection is enabled. This is due to MID servers affected by PRB1966095. Agents will make a REST call to the MID server asking for the number of agents the MID server has connected to it. The data call will also send back a complete state of the MID server which can be exceedingly large. Depending on how many agents/policies/checks are configured, this data structure that is serialized can exceed to maximum log sizing of the agent. This results in the agent attempting to write the JSON payload to the log at the INFO level during startup, having it exceed the configured maximum log length with one logging statement, causing an overflow and restart of the execution context. Since the restart will query the same MID server REST endpoint, the agent ends up in a continuous loop of archiving a previous zero length log file and generating a new zero length file due to the overflow. These files are generated at the millisecond level resulting in the file system of the agent endpoint becoming full.Steps to Reproduce - Set log-max-size in the acc.yml file to something really small- Start the agent with a MID server that queries for metadata affected by PRB1966095 (Z-)- Note the agent client collector will create zero length log files during startup at the millisecond level- Archives of the zero length log files accumulate until the disk is saturated.WorkaroundChoose one of the below fixes to prevent the accumulation of archived zero-length log files during agent startup: Update the MID server to include the fix to PRB1966095Disable auto-mid-selectionSet max-log-size in the acc.yml file to something greater than 10 MB (e.g., 100 MB)Set log-level in the acc.yml file to "warn"Related Problem: PRB1992666