Troubleshooting
Articles about how to solve the most common problems
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
MartinVolb
Dynatrace Participant
Dynatrace Participant

 

Summary

Dynatrace extensions can hit CPU or memory soft or hard limits enforced by the Extension Execution Controller (EEC). When this happens, tasks may be rejected or stopped, and monitoring configurations may fail to assign. This article explains how to diagnose, troubleshoot, and resolve these issues.

Problem description

Extensions running on an ActiveGate or OneAgent may exceed CPU or memory limits. When these limits are reached:

  • Extensions may fail to start or stop running.
  • Incoming tasks may be rejected.
  • Monitoring configurations may remain in Pending or Error state.

Common error messages include:

  • High system resources consumption detected for datasources (soft limit)
  • Rejecting configuration… Reason: soft limit reached
  • Hard limit reached, stopping extension tasks
  • Failed to assign monitoring configuration due to resource limits

These messages typically also include details explaining which limit was exceeded and by what margin.

Soft limits reject new tasks once usage exceeds the threshold. Hard limits stop tasks until consumption returns below the threshold. For background on limits and performance profiles, read Extension limits and performance profiles documentation.

Troubleshooting steps

Investigation steps

  • Open the Extensions app in your Dynatrace environment.
  • Locate the affected extension.
  • Check for configurations in Pending or Error state.
  • Select the error entry to view logs.
  • Inspect logs for limit breaches. Look for messages such as:
    • soft limit reached
    • hard limit reached
    • High memory usage detected
  • If logs are inconclusive, collect a Support Archive. This bundles logs and configuration details for troubleshooting.
    How to collect a Support Archive.

Log locations

The EEC writes limit-related errors on the host running the extension.

ActiveGate

Linux:   /var/lib/dynatrace/remotepluginmodule/agent/log/
Windows: C:\ProgramData\dynatrace\remotepluginmodule\agent\log\

OneAgent

Linux:   /var/log/dynatrace/remotepluginmodule/agent/log/
Windows: C:\ProgramData\dynatrace\remotepluginmodule\agent\log\

What to look for

  • Python-based extensions
    • python*.log: CPU/memory messages (soft limit reached, hard limit reached)
    • ruxitagent_extensionsmodule_XXX.log: lifecycle events, enforcement
    • EEC controller log: enforcement details
    • *.out and *.err: standard output and error
  • Non-Python datasources (SNMP, WMI, JMX, SQL)
    • Datasource-specific logs: runtime activity
    • remotepluginmodule.log: failures and enforcement
    • EEC controller log: enforcement decisions

Resolution

Dynatrace Managed

  • Use a dedicated ActiveGate for extensions (recommended best practice)
    A dedicated ActiveGate isolates extension workloads from ingestion, synthetic, and routing modules and improves stability under load.
    Dedicated performance profile documentation.
  • Switch to the High Limits performance profile (temporary mitigation)
    Temporarily increase CPU and memory headroom for extension execution by selecting High Limits under:
    Settings → Monitoring → Monitored technologies → [Your extension] → Performance profile
    Performance profile documentation.
  • Increase memory or CPU assigned to the ActiveGate
    If the host regularly approaches resource limits, allocate additional CPU or memory to the ActiveGate.
    ActiveGate configuration documentation.
  • Add another ActiveGate to the ActiveGate group where the monitoring configuration is assigned
    ActiveGate group documentation.

Dynatrace SaaS

Best practice summary

  • Use an ActiveGate configured exclusively for extensions and assign it to a specific ActiveGate group
    This ensures isolation of extension workloads, reducing resource contention and improving stability in SaaS environments.
    ActiveGate group documentation .
  • Use High Limits only as a temporary workaround
    Increasing limits can prevent short-term disruptions, but it may lead to resource exhaustion if used permanently. Always plan for scaling instead.
    Extension limits documentation .
  • Monitor logs for early warnings such as “High system resources consumption detected”
    Proactive log monitoring helps identify performance issues before they impact extension execution or ActiveGate stability.
    Extension troubleshooting and logs .
  • Add more ActiveGates to distribute extension workloads as environments grow
    Scaling out ActiveGates prevents overload and ensures predictable performance and failover capabilities.
    ActiveGate installation and scaling guide .

Configuration paths

Windows

C:\ProgramData\dynatrace\remotepluginmodule\agent\conf\extensionsuser.conf
C:\Program Files\dynatrace\remotepluginmodule\agent\conf\launcheruserconfig.conf
C:\Program Files\dynatrace\remotepluginmodule\agent\conf\custom.properties

Linux

/var/lib/dynatrace/remotepluginmodule/agent/conf/extensionsuser.conf
/var/lib/dynatrace/remotepluginmodule/agent/conf/launcheruserconfig.conf
/var/lib/dynatrace/remotepluginmodule/agent/conf/custom.properties

For details on overriding limits, read the EEC custom configuration documentation.

What’s next

Version history
Last update:
‎07 Apr 2026 12:19 PM
Updated by:
Comments
AntonPineiro
DynaMight Guru
DynaMight Guru

Thank you! :take_my_money:

AlainD
Observer

Hi Martin,

You speak about Scale horizontally by adding additional ActiveGates to the same ActiveGate group.

Good enough but I have a configuration with 3 AG with dedicated profile.

What I see is that I have only 1 AG who is working hard, with regularly soft limits alerts and 2 other doing almost nothing!

There is no load balancing occuring nor extension moved to another AG from same group.

For me seems  that Scale horizontally is useless. Any other solution to propose?

MartinVolb
Dynatrace Participant
Dynatrace Participant

@AlainD 

Hi Alain,

A controlled restart of the ActiveGate experiencing high load will allow traffic and processing to be redistributed to other ActiveGates.

Start/stop/restart ActiveGate – Describes how to restart an ActiveGate
mrc15816
Advisor

@AlainD you can force the endpoint to get evenly distributed, I believe it is called Task Bucket set to 1, which is available when configuring the end point.

Good luck

Thanks

Raj