44 comments

[ 5.2 ms ] story [ 99.3 ms ] thread
I was playing with tails and thought github just didn't allow tor usage. Github's service record is so good that I didn't even stop to consider this might be a technical problem with Github.
I have the GitHub status page bookmarked. Take from that what you will.
It appears to be down only for some regions.
Some pages (like the github.com landing and userprofile pages) including user sessions seem to be working from here. But not browsing a repository via http. Dunno if git pull is affected.
Not sure if if it's regional but from Switzerland it's been around 1 hour that any kind of connection to github goes in timeout.
The eclipse will destroy everything we hold dear! It started with the poorest failover system known to man!

GGit hub deserves every piece of criticism they get. How they can receive so much funding and help and still manage to go down so many times every year absolutely baffles me.

Of all the services I pay for, gitlab might be the best one in terms of bang for buck.

Reminder to avoid the github monoculture and move to other git providers as a backup.

Gitlab is open source and has auto-pull from github

Gitlab is always mentioned here as an alternative on most every github outage post. I am wondering if there are gitlab shills ready to fire in these events? There are many other git products as well.
The only major one that comes to mind is Bitbucket.

Gitlab comes up a lot because its idealistic goals resonate with a lot of the people here.

I'm not paid, but I am a proponent for a federated web.

I'm happy to host unlimited number of repos at my own expense[0] if it means people depending less on centralised hosting and allowing them to migrate away when they feel like it.

[0] https://git.drk.sc

They also have constant security issues. Some of these security issues make you doubt the quality of the code that Gitlab has...
Gitlab pays me $50 per upvote (and a bonus based on the number of comments and my comment rank)
Perhaps instead of assuming everyone else in the universe is a mercenary liar, you could look to other explanations?

I use Gitlab for some specific things. It's worth checking out. If only because it's nice to know what a migration strategy off githib would look like.

Also, it's really not hard to self host your own gitlab server.

I only push public works to Github, anything private is stored in my own Gitlab instance.

gogs is also open source, simpler and smaller than gitlab. also viable alternative i think.
Thank you for extending my todo-list :-)
Actually, if you're looking into Gogs, please look into Gitea instead. It was forked from Gogs because Gogs is done by a single maintainer, who just vanishes for months at a time every once in a while. (Which is not a bad thing per se, but this is precisely why Gitea is set up as a team effort.)
Good to know, I'll be sure to do just that. Much thanks.
Yeah I'm just going to submit a polite request to every variety of package/plugin/whatever manager that I use that somewhere somehow relies on GH to cut that shit out.
This is the second major outage of GH this summer. I'm fearing we should all become less dependent on this single service for bug reporting, merging, etc. and instead leverage the decentralized power of Git to host on a variety of platforms. That way if one is out, you just fall back to another.

Does anyone know of libraries that can sync at least the filed issues between GitHub and GitLab or Bitbucket in near real time? That's the main thing from GitHub that I'm truly missing when there's an outage.

Is this a major outage? They've been experiencing connectivity issues for ~20 mins. I've been able to connect just fine all the while, as well.

GitHub isn't 'down.'

And their status page has an unbalanced parenthesis [0]. I send their support an email about that. (There is also another on the messages page that I did include in the mail).

It's quite a silly thing to complain about, but I imagine other software engineers also get slightly stressed by seeing an unbalanced paren!

[0]: http://i.imgur.com/cypsCvx.png

(comment deleted)
I like to think that unbalanced parens accumulate in the soul, just like incomplete sneezes. Ones fate in the next life (perhaps starting equipment) is determined in part by the number of un-matched left parens hanging over ones head at the moment of death.
Is this an unclosed paren? ;)
Now I can't unsee them! What have you done! :(
I can't help but wonder if your username is a coincidence. :)
(comment deleted)
I bothers me how often I see only closing parentheses in Bash case statement patterns!
I am eating and I felt my stomach literally churn. I think it was a bit of an overreaction but considering they are a popular firm, not so much.
Queue obtuse comment about how this is no big deal because git is distributed source control, ignoring the fact that Github != git and that issues/PRs/etc are pretty important.
Followed up by the comment that we need to stop this monoculture that we are creating and go back to mailing lists.

Which is then countered by people arguing that GitHub was the greatest thing to happen to open source and has introduced so many more people to the OSS community. Perhaps throwing in that a self hosted solution will most likely have worse uptime than GitHub does (which spawns it's own sub-thread of SaaS vs self-hosted solutions)

Then someone says that a higher barrier to entry could help the ecosystem (while muttering something about how many JS libraries there are out there).

I could keep going, but it seems github is working again for me, so off to do some work!

> which spawns it's own sub-thread of SaaS vs self-hosted solutions

Our Github Enterprise instance sure has better uptime than Github itself. :)

also means that we cant composer install any project because even if the requirement is cached, each item is checked against github to see if the cached version should be invalidated :/ just took half an hour to complete an install...
Does github report RCAs for outages to the public? I'm wondering if these outages are related to their migration to kubernetes. Either way I'd love to learn what's going on and how they resolve these incidents. There's a lot to learn and I find these things super valuable.
(comment deleted)
Such an outage reminds us all how much we rely on GitHub just working. My 'flow' was suddenly interrupted because I couldn't continue my search for a specific PHP package. Luckily the workday is nearly over in the CET time zone so the impact on my productivity is minimal.
It's always been working from Poland from what I could tell. Is this possible?