APM becomes unstable and logs contain large number of fatal errors due to lack of RAM on the Server


 Product(s):APM Implementation and Performance Management
 Version(s):7.11
 Environment:N/A
 Area:N/A
 Subarea:N/A

Problem

User suddenly starts receiving increased amount of fatal errors (recorded in the error logs) but has not made any fundamental changes to APM (no hotfixes, updates or customizations applied to APM or the server) and is using the same functionality they normally would. when using regularly used APM functionality.

Solution

While fatal errors entries in the log file are often useful in describing the source of the failure, sometimes they result from a lack of system resources such as low/no available RAM or CPU. Typically a lack of RAM or CPU can produce a sudden increase in Fatal Error entries in the error log despite user using the APM product in their normal way. The fatal errors may appear to be spurious after initial investigation but the error logs suddenly start growing more rapidly and being over taken with this seemingly spurious errors.

The solution is to review the load and traffic on the server over several days to see if you are indeed experiencing high RAM or CPU usage and increase the amount of RAM and/or CPU as needed. Most users of recent APM releases 7.9 or later will find 8GB to be too limited and may wish to increase to 16GB or 32GB.

Over time as databases grow and APM develops more functionality, you may find your server resources being impacted more with each new APM release. As always, it is a good idea to review your system resource needs whenever you are performing an upgrade by reviewing the version's prerequisites.