Owing to an unscheduled reboot, caused by a kernel panic on the virtualisation host node (pve-1), a number of proxies associated with the Zimbra cluster went offline at approximately 10:15 this morning.
Engineers were notified and services were restored approximately 15-20 minutes later. At this time, we are monitoring the situation, which appears to have been caused by a hardware watchdog timer.
We would like to apologise for any inconvenience caused and are continuing to monitor the services to ensure they are operating at peak efficiency.
Note that the hardware watchdog did exactly what it is designed to do and where a kernel process hangs or anything untoward is detected, the server is intentionally rebooted to clear any problems. An updated kernel was also installed as a byproduct of the reboot, which we believe contains a patch for this issue, which last occurred on April 1st.
We are however continuing to monitor to ensure there are no additional problems.