OOM Issues Troubleshooting

OOM Killer Issues Troubleshooting

When an Out-of-Memory (OOM) killer alert is received by email, investigate which process was terminated, identify the process category, and apply the corresponding resolution.

The OOM killer is triggered when a container does not have enough available memory to continue operating normally. The notification email provides the process name that should be used as the starting point for troubleshooting.

On this page

Troubleshooting Workflow

1Open the OOM alert email

Review the notification sent by the platform and identify the process that was terminated.

2Copy the process name

Use the exact process name from the email when checking the troubleshooting documentation.

3Identify the process group

Determine whether the process belongs to Common Cases, Processes of High Risk, or Non-Leaking Processes.

4Apply the recommended resolution

Follow the instructions provided for that process and stack type to restore service and reduce the risk of another OOM event.

Process Categories

Common Cases Processes that are normally present in platform containers and can still be terminated by the OOM killer.
Processes of High Risk Processes that may indicate a memory leak, inefficient application behaviour, or a need for code and stack optimisation.
Non-Leaking Processes Processes that may be terminated by the OOM killer even though they are not the actual cause of excessive memory consumption.

Common Cases

This group contains processes that run by default in platform containers and may be selected by the OOM killer when the available memory is exhausted.

Use the process name from the OOM notification to locate the corresponding entry in the Common Cases troubleshooting guide.

Processes of High Risk

This category covers processes that may represent memory leaks or inefficient memory usage. The recommended actions depend on the software stack and programming language.

  • Review the stack-specific recommendations.
  • Check whether application-code optimisation is required.
  • Inspect the process-specific resolution provided in the related troubleshooting guide.
  • Apply corrective actions before returning the application to normal operation.
High-risk processes: Restarting the container alone may not prevent the problem from returning. The underlying application or configuration issue should be identified and corrected.

Non-Leaking Processes

These processes may be terminated during an OOM event even though they are not the root cause of the memory shortage.

The general recovery action for this category is to restart the container so the terminated processes can be restored.

After restarting the container, continue checking memory consumption to identify the process that actually caused the resource shortage.

What to Do Next

  1. Retrieve the killed process name from the email notification.
  2. Open the relevant process-category guide.
  3. Search for the exact process name.
  4. Follow the documented resolution for that process.
  5. Monitor the container after applying the fix.