Heap dump (OutOfMemory error) on BG when Trace Jobs is enabled, slow performance on BG

Products

• Clarity PPM SaaS

Issue/Introduction

We notice a performance degradation on BG, the jobs are running slow, and we also have outage(crash) with heap dump on OutOfMemory : Java heap space (OOM)

Resolution

In the heap dumps we can see that root cause of the performance issue with the BG service is that the SQL trace is enabled on the jobs. This creates a significant overhead on performance and can even crash the server when heavy jobs such as the Time Slicing are running. We do not recommend to have the tracing permanently enabled as it leads to issues.

 

1. To fix the problem - please connect as user with admin rights

2. Once you're connected, replace everything after action: with security.logs

Example:

 https://SERVERNAME:PORT/niku/nu#action:security.logs

3. Go to Logs - Edit Configuration

4. Uncheck Trace Jobs

5. The changes should be applied automatically if Detect Changes Automatically is checked. If not, a services restart is required.