Best practice - concerning undebug all
When using debugging commands on a Cisco device, it is vital to ensure that the chosen commands do not overwhelm the resources of the device, especially if it is a device serving a production network.
If too many debugs are issued, or if there are many events on a device taking place that trigger debug messages, the CPU and/or memory may be quickly overwhelmed, causing the device to delay in performing its function or even to fail completely.
Some precautions that can be taken to avoid such a situation is to ensure that the undebug all command is issued regularly to make sure that no forgotten debugs are still running.
It is possible to configure a Cisco IOS device to periodically issue the undebug all command. This can be done using the Kron task scheduler, for those devices that support it.