Looking at the logs from Chris Gregan, the latest 'since' time in the juju_status.yaml is 11:01, with most of the 'since' times occuring between 10:50 and 10:55 - if the juju_status.yaml was captured at the same time crashdump was run (as is usually the case), it would indicate a controller problem - multiple units were still running update-status well beyond 11:01, e.g.
Looking at the logs from Chris Gregan, the latest 'since' time in the juju_status.yaml is 11:01, with most of the 'since' times occuring between 10:50 and 10:55 - if the juju_status.yaml was captured at the same time crashdump was run (as is usually the case), it would indicate a controller problem - multiple units were still running update-status well beyond 11:01, e.g.
https:/ /pastebin. canonical. com/p/vKf6pMKzN k/
I assume something broke on the controller node at approx. 11:01. I guess we need to see the controller logs if this happens again?