Hacker News Reader

So one outage in 10 years has you this concerned and worried that you're considering moving? Wow!

It’s not the outage, it’s the response, according to GP

If there’s an average of one or two outages a decade, I can forgive an unpracticed protocol.

It seems like they started replying on Twitter https://twitter.com/Fastmail/status/1674788776202350593

No new updates, just that they are still working on this issue.

That's fair, but I don't think either Google or Microsoft (being the two alternatives proposed) are terribly forward with their downtime notifications either.

I work in email security. If we even delay some emails for 5 minutes+ we're going to see significant customer response. Systems are designed accordingly.

Is email really the right medium for that real-time communication? Being async in nature? What about spam filtering etc?

Not really. email is more of 'we'll get there eventually'.

We get to spend an average of 500ms with each email in terms of scanning. Obviously some things may spend a much longer time but we keep it to the minimum.

Absolutely email is not intended for time sensitive Comms but absolutely it is used for them.

It is more about how they handle it right now. I always only had the best to say of FastMail team. Great support, great team. And personally I always felt like I did not pay enough for using their service.

But it concerns me that we are in silent for 7 hours.

If you follow the Sysadmin subreddit, Microsoft routinely has long outages. Last week (or the week before) email was completely down, the admin panel was unusable, etc. Fastmail has been FAR more reliable then Microsoft. Yes, the Fastmail team should be responsive though. At least with Microsoft they internal comms are very leaky with this sort of thing so details are usually public long before it's announced.

> so not sure if some other mail was bounced

SMTP is designed to be resilient to outages. The sending MTA is supposed to queue messages if it can't reach the destination MTA. It should only bounce them after a significant outage (days, typically).

https://www.rfc-editor.org/rfc/rfc5321#section-4.5.4

In the last few years, GitHub reached a point where it seems like it's partly down every week. Even Facebook is down from time to time. WhatsApp too.

One outage in 10 years sounds like a good reason to not move away, if you ask me.

Apple Mail's not quite as customizable, indeed. Maybe close though.

I made a table comparing different email services on the MacPowerUsers forum a while back. It's in need of updating but might be handy: https://talk.macpowerusers.com/t/icloud-mail-fastmail-gmail-...

Thank you, this is useful. And actually brings the most important concern: the support. If something will go wrong, you can talk to real people behind FastMail. In case of Apple - not sure what is going to happen.

In a similar situation/location. But can't access it through my home internet, but my phone is just fine if I turn off wifi. This is funny because I was still getting push notifications, I just couldn't read them with my phone connected to the home router.

They've been brilliant otherwise. Able to import everything from gmail very easily.

> But can't access it through my home internet, but my phone is just fine if I turn off wifi.

Thanks for the tip. I can't access it via my home provider, but T-Mobile works. Interestingly enough, I'm getting new email notifications on my home provider; I just can't read the mail through it. They're probably using an unaffected Google service for notifications.

Looked at the Apple Mail/iCloud Mail - they support only up to 5 custom domains. Which I guess you can workaround by creating separate iCloud accounts and pay for iCloud+

> they mentioned "We're in close communication with our provider", which probably means it is not their fault, but the ISP.

Shouldn't a system this big be redundant?

> , so not sure if some other mail was bounced.

99% of outages on mail are supported totally by the protocol layer. It's a beautiful thing that if mail isn't delivered, the sender will just try again layer. It makes restarting services in a mail company much, uh, less risky. You won't miss any messages.

Other mail systems should hopefully be configured for retry up to a few days.

So I wouldn't think the mails should bounce unless someone used a really short retry period.

"Retries continue until the message is transmitted or the sender gives up; the give-up time generally needs to be at least 4-5 days." from RFC 5321