Table of Contents:
On a site configured with high availability (HA), why do the servers have different heap utilization profiles?
Why is my heap profile different after upgrading to 20.1?
Appian uses the G1 Java garbage collector beginning in Appian 20.1. Thresholds dictating garbage collection frequency and volume have changed. This is due to the change in garbage collection strategy in order to optimize for both system stability and performance. Therefore, Appian Cloud sites may display new heap utilization patterns despite unchanged usage.
What should I expect when looking at Cloud Insights?
Heap usage may continue to fluctuate even during periods of no system activity due to Appian system components running in the background. It is common to see fluctuations as high as 80%. Below is an example of a common heap utilization pattern while using the G1 garbage collector:
Each server in a HA configuration manages heap independently, so it is not uncommon for servers to differ in their usage profiles. This is especially common on startup, and is not indicative of an issue.
This article applies to Appian 20.1 and later, as well as all versions of Appian using the G1 garbage collector.
Last Reviewed: September 2021