tree 3dff4cdf264bed17e66f7aed2c8085b67738104d
parent 86a714d6e81bb524dc59fda7baa63b45e7180489
author Serge Bazanski <serge@monogon.tech> 1681817328 +0200
committer Serge Bazanski <serge@monogon.tech> 1681912501 +0000

cloud/shepherd/equinix: implement recoverer

This implements basic recovery functionality for 'stuck' agents. The
shepherd will notice machines with a agent that either never sent a
heartbeat, or stopped sending heartbeats, and will remove their agent
started tags and reboot the machine. Then, the main agent start logic
should kick in again.

More complex recovery flows can be implemented later, this will do for
now.

Change-Id: I2c1b0d0465e4e302cdecce950a041581c2dc8548
Reviewed-on: https://review.monogon.dev/c/monogon/+/1560
Tested-by: Jenkins CI
Reviewed-by: Tim Windelschmidt <tim@monogon.tech>