The supervisor should be able to detect restart loops, and exponentially backoff

balena-os / balena-supervisor

Balena Supervisor: balena's agent on devices.

https://balena.io

Other

149 stars 63 forks source link

The supervisor should be able to detect restart loops, and exponentially backoff #961

Open CameronDiver opened 5 years ago

CameronDiver commented 5 years ago

This will aid in both debugging and in the cases such as the container taking down the network, will provide a time in which the supervisor can update itself out of the situation.

As an extension we can also track this metric over reboots, in case the faulty container is somehow causing the device to reboot.

jellyfish-bot commented 4 years ago

[gelbal] This issue has attached support thread https://jel.ly.fish/99b9dbc2-2158-4ec2-af9e-8aee6003b2c1

20k-ultra commented 4 years ago

One case to prevent reboot loops is checking that the device has enough disk space for the operation that it is going to try and perform which is seen in the above JF link.