This is article explains how to deal with Unaccounted memory computation on Brackla Devices 

On Dec 19th 2023:

root@EVO_Lab> show chassis routing-engine no-forwarding

Routing Engine status:
Slot 0:
Current state Master
Election priority Master (default)
Temperature 32 degrees C / 89 degrees F
CPU temperature 46 degrees C / 114 degrees F
DRAM 13219 MB (65536 MB installed)
Memory utilization 79 percent  <<<<<<<<<<<<< 

On Jan 19th 2024:
root@EVO_Lab> show chassis routing-engine no-forwarding

Routing Engine status:
Slot 0:
Current state Master
Election priority Master (default)
Temperature 31 degrees C / 87 degrees F
CPU temperature 45 degrees C / 113 degrees F
DRAM 35411 MB (65536 MB installed)
Memory utilization 45 percent  <<<<<<<<<<<<<<<<<<<<<

The zookeeper snapshots/transaction logs in ramfs(filesystem) causes incremental usage of memory.
Currently the disk manager(diskmgmt) cleans up the logs/snapshots when the usage reaches 85%. 

Zookeeper logging was counted part of memory. 
The fix provided to clean up the logs periodically once in 8 hours so memory will not increase continuously.

There is no functional impact with this issue. 
However, these memory utilization cycles could interfere with device memory utilization monitoring hence enhancement of proactive clean up of memory every 8 hours was proposed/added.

This issue can be observed when "show system memory" and "show chassis routing-engine no-forwarding"  outputs are compared over a period of time 
or 
when the SNMP OID jnxOperatingBuffer is polled to monitor device memory utilization.

This issue might be seen if the following conditions are met:
* On all Junos Evolved platforms
* Anything that uses zookeeper extensively can cause this issue.  For example: AE(aggregate ethernet) interfaces that flap often etc.