MemoryCritical¶
| Severity | critical |
| Fires when | Memory usage is above 85% of the memory limit for 2 hours |
What it means¶
The service uses more than 85% of its memory limit, and this has lasted for 2 hours. The memory limit comes from the compute plan of the instance.
What happens if you do nothing¶
If the service reaches the memory limit, the platform stops the container and starts it again. The service is not available during the restart. This can happen again and again while the load stays high.
What to do¶
- Check the current memory usage. Open the Technical Details on the instance detail page. The Memory column in the Pods table shows the usage against the limit. See Monitor an Instance.
- Check the logs for out of memory errors and for restarts. See View Logs.
- Find out why the memory usage went up. Common reasons:
- More traffic or more users than before.
- A new version of your application that uses the service in a different way.
- A recent change to the instance configuration. The audit log lists all changes.
- Reduce the load, if the high usage is not expected. See the steps for your service below.
- If the load is expected, move the instance to a larger compute plan. See Scaling. A compute plan change can cause a short restart.