View Single Post
Old 8 Aug 2018, 03:05 PM   #44
BritTim
The "e" in e-mail
 
Join Date: May 2003
Location: mostly in Thailand
Posts: 3,095
Quote:
Originally Posted by FredOnline View Post
From the FastMail Blog:

The full story on the FastMail outage on August 1st: what happened, and what we'll be doing to prevent it from happening again:

https://fastmail.blog/2018/08/08/fas...ge-2018-08-01/
This is generally a pretty clear explanation. However, there are two issues I can see
  1. Unexplained is why (as reported by Pingdom) the secondary MX server through Seattle went down for about an hour. There ought to be complete incoming network isolation between the different MX servers to ensure appropriate levels of availability. This is significant as the clean solution for clients to set up backup email through another service relies on high availability of the MX servers.
  2. Although some mitigation is being put in place, and NYI has proved reliable in the past, the fact remains that FastMail is going to remain vulnerable to NYI as a single point of failure. IMHO this means that those who would be severely hurt by email outages need to set up email backup. FastMail does make this easy (as long as at least one MX server remains available) but the need is unfortunate. Also, based on the outage last week, it seems we cannot rely on an available MX server.
I do not intend to take precipitate actions. I have not introduced new clients to FastMail for some years. The question is whether in switching anyone away from FastMail is prudent or outweighed by the disruption involved.
BritTim is offline   Reply With Quote