Troubleshooting High Resource Usage from beam.smp

What is the issue?

The beam.smp process on some Opsview collectors may use a high amount of CPU or memory, which can cause the system load average to increase. This is not caused by a lack of CPU, memory, or disk resources.

When does this happen?

This usually happens when the beam.smp process has been running for a long time.

Resolution

## Stop the Opsview components executor & scheduler:
/opt/opsview/watchdog/bin/opsview-monit stop opsview-executor
/opt/opsview/watchdog/bin/opsview-monit stop opsview-scheduler

## Terminate beam.smp process
pkill -f beam.smp

## Restart the Opsview components
/opt/opsview/watchdog/bin/opsview-monit start opsview-executor
/opt/opsview/watchdog/bin/opsview-monit start opsview-scheduler

## This helps reset the process and reduce CPU and memory usage.

What should I do after fixing it?

After running these commands, monitor the system for a few minutes or hours to confirm that CPU and memory usage remain stable. You can also recheck the service checks to update their status. Your local checks should return to their normal state, and the ‘Unix Load Average’ check should report an ‘OK’ status.

Is this a permanent fix?

No. This is a temporary workaround. If the process runs for long periods, the issue may return.

For further troubleshooting:

datastore_notifications_housekeep — What it does and why it matters?

Note: Instructions for installing the support script can be found here: Opsview - Install support scripts – Support - ITRS Group

Overview Copied

datastore_notifications_housekeep is a housekeeping script used to clean up old notification data from the opsview-logs database stored in opsview-datastore.

As of Opsview version 6.7.x, this database does not have automatic housekeeping enabled. On busy monitoring systems that generate a high volume of notifications, this can lead to the database growing continuously and consuming significant disk space over time.

This script helps prevent that by removing notification records older than a specified number of days.

/opt/opsview/supportscripts/bin/datastore_notifications_housekeep [-d <days>] [-s <sleep>] [-v]

## remove anything over 6 months old
/opt/opsview/supportscripts/bin/datastore_notifications_housekeep -d 180

Options Copied

Files Copied

Important Notes Copied

• This operation permanently deletes historical notification data.
• Always ensure you have backups if notification history is required for audits or reporting.

When should I contact support?

• If high CPU or memory usage continues after restarting the services, or if it happens frequently.