Looks to be their authentication layer since existing connections kept working even when new connections couldn't be made.

It always seems to be auth, the gatekeeper is the single point of failure in many systems it seems.

Yeah, hard to see what else would take down every region at once.

Yikes, there's not separating availability zones, then there's not separating entire regions and continents.

They may have a single point of failure shared in a sub-sub-sub-service. Or, they may have pushed not-so-great code into all regions/zones and whatnot.

AWS IAM is no different.

https://news.ycombinator.com/item?id=24498678

That post mentions us-east-1, so it's not really the same (limited to region).

I appreciate their honest status page.

In contrast to Amazon/Cloudflare’s “some users may experience delays” when in fact everything is down for a whole continent.

The "some users" wording comes from the fact that at any point in time only a small proportion of users are trying to use the service.

Anyone not using the service when it is down isn't experiencing a service outage.

Wow, Snowflake can do incident notification via webhooks! Pretty cool.

I wonder if that is what is causing the Slack issues

Whats happening with the slack now ?

Was slow to load messages for quite a bit. Seems fixed though

This status page... the lack of updates... it's like watching snow melt

I came here expecting global warming jokes.

It would make sense if the outage was constrained to US-West, given the heatwave and all.

Global outages make a lot of people feel good about not bothering to set up multi-region deployments though.