High Memory Usage Due to Cortex Telemetry Backlog

[cvrsilva]

I'm experiencing high memory usage issues on some endpoints, and about 99% of it appears to be caused by Cortex.

Why does this happen?
Cortex consumes telemetry in real time. However, if it fails to send the telemetry from the machine to the Cortex gateway (for various reasons), it accumulates the data locally and then sends it all at once. This behavior is common among security agents that operate in real time.

Current Solutions (and Their Limitations):

• Restarting the machine – Sometimes not feasible.

• Pausing Cortex on the machine – Requires time, as it depends on the machine owner contacting the security team.

Has anyone experienced this before? How did you resolve it?
Is it possible for a local administrator to disable the Cortex service on the machine directly?

[aspatil]

Hello @cvrsilva ,

Please check the Network Stability: Ensure the affected endpoint has a stable connection to avoid telemetry backlog. Also agents are running on latest version.

If the issue persists, you may need to engage Palo Alto Networks support for deeper investigation into resource utilization.

If you feel this has answered your query, please let us know by clicking like and  on "mark this as a Solution". Thank you.

[cvrsilva]

There were no holes in the network. The agents are update version.

[cvrsilva]

The high CPU usage issue sometimes occurs on Linux endpoints and occasionally on Windows Server.
Are there any best practices to prevent this?