Somehow I'd missed this when I added the state diagram originally.

This will enable debugging/introspection on whether a service is
failing and being auto-restarted. You can introspect it with the
changes API or the "pebble changes" and "pebble tasks" CLI commands.

The kind (type) of the changes is "recover" and the kind of the tasks
is "restart".

Example output while I service is begin recovered (change 3 is not yet
ready):

$ pebble changes
ID   Status  Spawn                Ready                Summary
1    Done    today at 16:08 NZST  today at 16:08 NZST  Autostart service "test2"
2    Done    today at 16:08 NZST  today at 16:09 NZST  Recover service "test2"
3    Done    today at 16:09 NZST  -                    Recover service "test2"

After it's recovered the change is marked done/Ready:

$ pebble changes
ID   Status  Spawn                Ready                Summary
1    Done    today at 16:08 NZST  today at 16:08 NZST  Autostart service "test2"
2    Done    today at 16:08 NZST  today at 16:09 NZST  Recover service "test2"
3    Done    today at 16:09 NZST  today at 16:10 NZST  Recover service "test2"

$ pebble tasks 2
Status  Spawn                Ready                Summary
Done    today at 16:08 NZST  today at 16:08 NZST  Restart service "test2"
Done    today at 16:08 NZST  today at 16:08 NZST  Restart service "test2"
Done    today at 16:09 NZST  today at 16:09 NZST  Restart service "test2"

$ pebble tasks 3
Status  Spawn                Ready                Summary
Done    today at 16:09 NZST  today at 16:09 NZST  Restart service "test2"

Fixes canonical#104

See canonical#119