HN Explorer

Ask HN: Why did both “fastmail is down” posts get flagged?

3 points by comboy ↗ HN
HN is always full of positive opinions about Fastmail. Why would posts about its downtime, which seem relevant to many HN users, get flagged? Downtime seemed to have shown some important aspects about their infrastructure (which I mentioned in my comment on one of these posts)

6 comments

[ 2.1 ms ] story [ 24.6 ms ] thread
alphabettsy
There seem to be lots of emotional flaggers, that's about all I've gathered. Maybe someone could elaborate further.

I personally am seeing a delay in receiving(1-3 mins), but not sending mail, both via IMAP. I have not checked webmail. Assuming this is a blip, FM certinaly will still continue to be my choice since this would be the first outage I've noticed and I love most everything about the service.

Status page shows degraded still: https://www.fastmailstatus.com

greenyoda
Because they're off-topic. HN Guidelines say:

"On-Topic: Anything that good hackers would find interesting."[1]

Why is a brief interruption in Fastmail or Twitter or Netflix interesting? These things happen all the time. As a Fastmail user, I noticed that Fastmail was down, and then I looked at their status page to confirm they were really down and the problem wasn't on my end. Then I waited five minutes, and they came back up. Seeing a post about it on HN doesn't help me at all. And someone who is not a Fastmail user wouldn't care in the least.

They said they lost network connectivity to a couple of racks. What interesting things can we learn about their infrastructure from this? Their back-end IMAP server got disconnected and their web servers stayed up. It's a completely generic problem that can happen in any datacenter.

[1] https://news.ycombinator.com/newsguidelines.html

detaro
Agreed. If there is something interesting to learn, they'll hopefully write about it later. That's then something we can have a meaningful discussion about.
alphabettsy
I'm inclined to agree with you, but there was a massive thread last time S3 had an outage.
comboy
Their website was also down.

Later on the website there was a static message: "I'm sorry, the server your email is on is currently down. We apologise for the inconvenience; any email sent to you during this time is being queued by another server and will be delivered as soon as the server is working again." so their mail servers were also down, not just the interface.

And the Twitter message says "We've just lost network to a couple of racks,"

So to me it's interesting that it seems their whole infrastructure depends on these couple racks in a single place. Many people rely on e-mail for their work, and DNS updates are slow, so it's not easy to quickly start receiving e-mail elsewhere if anything happens to that place.