|
My initial survey of the Pingdom data suggests that the problem is very similar to the last complete outage. However, it was resolved more quickly. Once again, both MX servers were unavailable, including the secondary MX server (in Seattle). Both MX servers were mostly unavailable for about 35 minutes though the Seattle server was reachable for a brief period in the middle.
The practical consequence for me is not severe because it occurred on a Sunday. On a regular working day, it would have implied some mail delivery (including to non FastMail backup mail servers) was delayed for an hour (the typical time between an initial attempt to deliver the messages and the sending server trying again). This can be serious for a couple of my business clients.
Essentially, if the secondary MX server in Seattle is unavailable at the same time as the primary MX server, it makes the secondary MX server absolutely worthless. IMHO, FastMail should assign the highest priority to eliminating the shared failure point that results in simultaneous inability to access both MX servers.
|