GitHub was down for 5-10 minutes today. Two people got upvoted for reporting it. That makes no sense to me. Whenever GitHub goes down for more than 30-40 minutes, then yes, it's a serious disruption, anything else pollutes people's RSS feeds unnecessarily. Please consider this next time. Thanks.
Why report it at all? Seriously. If I'm affected, I'll find out. If I'm not affected, why should you tell me?<p>It seems that every time a well-known service goes down, for however long, for whatever reason, instantly there's a flurry of posts here making sure everyone knows something that they either already know, or won't care about.<p>I'd really like that to stop, although perhaps I'm just a curmudgeonly old grey-beard.
This post is worse than what it derides, of course. It fits well within the category of things that pollute people's RSS feeds unnecessarily.<p>This is hacker news, and many of the hackers here use GitHub so when it goes down they might be spinning their wheels. If it is down for less than three minutes it probably won't make it off the <i>new</i> page. I think your arbitrary <i>30-40 minutes</i> isn't better than what became the norm based on user behavior on HN. Why do you, Toshio, think you know with a high level of precision, how long GitHub needs to be down for it to be relevant to HN?
We launched our status board for exactly this reason, to get notified of random disruptions (specifically API disruptions). You can even chose your flavor, we support email, SMS, IM, webhooks, etc... It sure beats relying on HN or random failures for disruption notification.<p><a href="https://zapier.com/status/" rel="nofollow">https://zapier.com/status/</a>
Based on the speed with which the post received votes, I suspect most of the "votes" came from people trying to submit the news that had already been submitted.<p>I agree with you, OP, but it seems there were a lot of people jumping on that bandwagon.
I find twitter a much better forum for finding if a service has gone down. We experienced Lovefilm going down the other night and could easily confirm it was a server issue via twitter.<p>Of course, HN is a great place for people to discuss how to avoid a productivity disruption when such services go down.
I dont know why so many people always post it.
Please use Twitter for that.<p>Or just look at<p><a href="https://status.github.com/" rel="nofollow">https://status.github.com/</a>
I thought it was pretty serious. Github is no longer a tiny company and when so many people rely on you for everything a 5 min outage becomes significant. E.g. I was not able to deploy to my server a few mins ago due to this outage. This makes me question the decision to use github going forward.
I wish I could upvote you more than once. Actually, I was about to post something really similar.<p>Thanks. Github, Gmail, Google Docs, anything: more often than not, the moment I click the link they're working perfectly.
A billion times this!<p>Twice in the last 4 days, 2 stories to hit the front page in a row have been "ZONOES GITHUB DOWN!!!" when it's been a small blip.
Not sure HN is the best place to report it, however working at a company that extensively uses github on a large scale, for the business I think even an outage of 5-10 minutes counts as serious. I'm amazed that people seem to find downtime of 'cloudy' services more acceptable somehow. We have similar issues when JIRA goes down that has a significant impact on overall productivity.
I would respectfully disagree. One of the ways that I assess whether to use a service is based on its reliability. There is not a great way to look at historical records for when a particular service was down to understand how this might impact me. Searching HN has been very useful to assess other people's experiences with services, so why not for uptime too?
Come on, GitHub is used for serious work for several big companies. One hour downtime is an eternity, I'm quite happy with reporting 10 minute downtimes, as most of my workflow involves github, every step of the development process (ticketing system, dev vm's, CI server, capistrano, etc...) all at some point connect with GitHub.
I recommend monitoring third-party services automatically all the time, e.g. we do it using <a href="https://github.com/alphasights/open_nut" rel="nofollow">https://github.com/alphasights/open_nut</a>.
I do agree reporting outages on HN is totally irrelevant
Agreed, this just pollutes the eco-system around. If someone needs to know and gets impacted - they can check the status page themselves. Wasted 5 mins reading it!!
I agree.<p>But to play devil's advocate, in aggregate, the reports of the outages, even the small ones, could be useful to someone deciding whether or not to use GitHub.
I'm going to be that snarky fellow who submits a post and turns this thread into an ultimate meta-inception:<p>"Beg HN: Please only beg about serious issues (250upvotes+)"<p>Now upvote me for making clever comments about infinite-regressions, which all hackers are obviously interested in; or downvote me because I failed to add "</sarcasm>" to my comment - but wait, I just did, which would then cause an alligator paradox! (woohoo, now you'll want to upvote me because I mentioned paradoxes - but wait isn't that a paradox to upvote me for... nevermind)<p><i>But obviously you now want to downvote me because it is apparent I'm procrastinating and wasting time on HN and have nothing better to do. But wait, oh snap - now you want to upvote me because I'm writing satire about people who write about infinite regressions... which, wait, hold on, would mean that I'm not -- nope, nevermind. I'm shutting up here, because I'm sure you could figure out what my next 100 paragraphs will be, which means I don't even need to write it.<p>But gasp, I just did write th -- </i><i>this author was shot dead</i>* (then who wrote that he was shot dead? Obviously it was only a flesh woun -- <i></i>this author was `Rabbit of Caerbannog`ed<i></i>)