Slack outage: Connectivity issues | Hacker News Reader

Slack outage: Connectivity issues

status.slack.com

68 pointssamaysharma39 comments

Open articleView on HN

If only we had distributed realtime communication tech. Walled gardens FTW!

IRC

Does a non-Slack IRC channel go down less frequently?

IRC networks are distributed, so when one server dies you can connect to another.

So yes, IRC channels have better availability than Slack.

Generally yes, it's old and stable, and it's not centralized so not everyone goes down at once. When you host your own you often get to choose the downtime as well, updates get run at midnight when no one is working, not at 10AM on a work day.

A service the scale of slack also runs into a whole host of scaling issues that a little intranet IRC client will never have.

Which distributed realtime communication tech do you propose? IRC? Matrix?

Other threads:

https://news.ycombinator.com/item?id=15597431

https://news.ycombinator.com/item?id=15597387

https://news.ycombinator.com/item?id=15597382

https://news.ycombinator.com/item?id=15597382 looks like the original, but it's flagged, I guess?

What does that mean?

Slack's status page seems to have been in and out whilst this has been happening too (and sometimes the page loads but assets fail):

https://tkcs.in/461W2z0c1M3z

It's concerning/odd for their status page to be affected by whatever infrastructure issue they're having.

Same thing happened to Amazon when they had that big outage recently.

For reference. S3 went down. The amazon status page stored their icons on S3.

It might be on different infrastructure - just not stuff prepared for the onslaught of it being checked when there's a global issue.

The errors have been 500 responses, returned quickly. That suggests something other than high load (though it doesn’t guarantee it by any means).

Oh good. I was going nuts trying to figure out why I suddenly couldn't communicate with my coworkers and Slack kept saying I didn't have internet when I clearly had internet.

Ah, the downside of relying on a third party for critical communication infrastructure when you're a full remote company...

I guess people will actually have to get some work done

Collaboration with your co-workers is work.

Lol keep telling yourself that.

If you get paid to do it then it’s arguably more work than is toiling at side projects or posting snarky comments on hackernews.

As a professional you can't just go off in a corner and produce something (no matter how fun it is). I learned this the hard way as a software engineer. In fact the hardest part of software engineering is collaborating with humans and integrating their ideas and requirements into what you're building.

Helping other people, not producing output or profit, is the fundamental human enterprise. Other people exist and need help: It's the only reason you have a job. Amazing how many people miss this and think their job is to "be good at X" instead of "help other people with X".

It's important, but IMO it isn't the hardest part. Most people can do the collaboration and communication, but can't write code.

True, if you are talking about general population. But among those who can code in my experience the biggest obstacle at work is being able to communicate and collaborate efficiently.

Actually, the other 354 days a year when the site worked fine were work too. Ensuring the site is available and working smoothly is real work.

Grandparent comment is pretty clearly referring to Slack users, not employees.

Slacks status page also returns unminified css and four trackers (including twitter advertising).

> unminified css

Curious: does that even matter, assuming response body is gzipped?

Posts about slack seem to keep getting flagged away :/

We need at most one story about this outage on the front page at a time. Dupes should be flagged.

There wouldn't be dupes if the first submission hadn't been flagged away.

There are a two on the home page at the moment, but originally The first one was pushed right off the page.

At this time that post has 116 points in 35 minutes.

This one is the dupe, of https://news.ycombinator.com/item?id=15597382

Sigh... 10 minutes of cursing trying to troubleshoot the LTE connection on my phone. Sadly Slack has become a fairly critical part of our communication.

I've been having degraded services all morning, guess it finally kicked the can and went out.

It's been over 30 minutes - how does a chat app go down for this long?

It's built on human-constructed infrastructure or automated infrastructure designed by humans

It was bound to happen eventually. The scale of their application is mind-blowing.

Building things is hard, mistakes happen, and that's totally okay.

Well it's up again.

part of the screenhero rollout perhaps?

I just quit a call where we were using screen hero. It worked great. Then I noticed my status said I was still on a call so I thought I'd restart slack. It didn't really want to quit, so I killed it. Sorry everyone.

Reply on news.ycombinator.com