What is the performance impact on Foglight agents when the FMS is either down or unreachable (e.g. because of a network problem)?
See KB101111 "FglAM configuration to store more messages in case of FMS communication loss"
The impact on agents will be minimal. There will be a slight overhead due to their retry attempts to the FMS. For the agents that use SPID or FglAM (e.g. Operating System agents) you may also notice some increase in CPU and memory as they cache data while the FMS is unreachable. Also there will be some overhead when these agents send that cached data on to the FMS when it is contactable again. Other agents which do not use SPID or FglAM (e.g. the J2EE agent) does not cache data so you will not see the impacts caused such caching.
The specifics of what impact is caused will depend on the exact setup and configuration of the monitored environment.
However please note here have been some instances where FglAM will grab large amounts of memory while caching, and not releasing the memory for several hours. If you are affected by this particular issue please contact Quest Support.