NEXUS-12677, there are no obvious messages when the proxy repository not found cache has been exceeded or evictions have occurred.
The only real way to tell ( other than actively monitoring the JMX exposed mbeans for ehcache) is a debug message detailing the response for the item you've requested and then you would have to know it has been previously been requested to know that an eviction had occurred (really not useful in a support or diagnostic context).
Trying to force ehcache to log evictions using only the logging config in Nexus does not yield anything useful.
A possible solution would be to extend Cache Event Listener and use that to keep an eye on specific caches? Another idea is a thread which simply inspects the stats which are already being gathered on a regular interval, although has tradeoffs in not being real time.
- actively monitor cache statistics for every proxy repository not found cache ( negative-cache )
- at regular intervals determine the effectiveness of each not found cache based on stats ( algorithm TBD )
- in situations where the caches size is not being effective ( ie. too many misses on a too frequent basis ) log a message indicating this inside nexus.log at WARN level