Updated DomainMonster finally resolved problems with its hosted email service on Tuesday, more than two weeks after they first began on 21 December.
Reg readers affected by the problem were eventually told that a data centre networking issue was to blame for the extended outage. This remains unconfirmed since neither DomainMonster, a UK provider of domain name service, nor its owner Host Europe have provided a substantive response to our queries at the time of publication.
Inbound messages were worst affected in the outage, during which email servers hosted by DomainMonster "barely functioned form 21 December until Tuesday afternoon", according to Reg reader Phil.
"Between December 21 and yesterday I received a fraction of normal email volumes (my guess would be a quarter to a third), all delayed," he said. "I'm sure that some incoming emails were bounced back to their senders (I sent some test emails) but some did get through. I don't think that outbound emails were affected."
Both Phil and Bill, another reader who got in touch, complained about a lack of information from DomainMonster throughout the degraded service.
DomainMonster's status page – which currently suggests everything's A-OK – was often unavailable during the outage. Replies to support messages were made but these often took days to appear, possibly because festive-period support staff were overstretched.
Reg reader Bill pointed out that the latest extended outage of DomainMonster's hosted email service had a similar effect on users, if not necessarily been down to the same cause, as when the service went TITSUP for around a week back in 2014. Bill described the latest problems as a "repetition of 2014 events", covered in a Reg article at the time here.
The latest outage resulted in multiple frustrated complaints directed towards DomainMonster's hard-pressed social media support team via Twitter.
Copies of correspondence forwarded to El Reg provide a timeline of sorts. A day after the outage begun on 22 December, DomainMonster was telling customers it was on top of the issue but there was still a "large backlog of emails that need to be delivered".
By 29 December, and without any resolution in sight, the firm began blaming DNS problems for the snafu. "It seems the domain was using the old email DNS records," its support staff told our tipster Phil. "We moved across to new records recently, however the old set should still work but recently we have had reports of issues caused by the old set."
Only by the start of the new year had DomainMonster isolated the issue, which it blamed on a networking problem. "We believe that this is a network issue and our data centre team are looking into this," support staff said. "This issue seems to be preventing our external and internal email services from talking to each other and this has caused a massive backlog of email."
Critics blame poor network architecture choices for creating the problem in the first place. "Two of the three listed @domainmonster DNS servers are still down and all three sit behind the same router in London," said one such critic, Alastair Mackinlay, in a Twitter update on Sunday. ®
Updated to add
A spokesperson for DomainMonster has been in touch to say: "We are always working hard to ensure our system can be the best it can possibly be for our customers, and this mail connectivity issue was the result of one of our planned system upgrades. We are really sorry for any inconvenience caused, as we know email is one of the most important assets to our customers. We would like to assure our customers the issue has now been resolved."
Sponsored: Ransomware has gone nuclear