Hacker News Reader

Comment by MuffinFlavored | Hacker News Reader

Does anybody else feel GH has a much higher than normal incident rate?

Unless they under-report, we shouldn't need to feel anything; we can just check https://www.githubstatus.com/history and objectively look at frequency. Of course, some services do under-report; dunno if GH does or not.

The overwhelming majority seem to under-report these days. I can’t tell you how many conversations I’ve had that go like this:

—

Me: We’re having issues on our site, and we’ve narrowed it down to your service. Can you confirm that you’ve had a partial outage for the last three hours?

Them: Yes, about 25% of API calls are failing.

Me: Why is your status page green?

Them: Our team has decided it doesn’t warrant a status update.

—

When this happens, we tend to switch services providers fairly quickly. If you’re not going to be honest with us, we’re stuck wasting our time troubleshooting. We don’t care if you have occasional outages—it happens; that’s why we have fallback providers. But you need to tell us when you’re down.

tl;dr: Your status page is not an advertisement for your “amazing” uptime. It’s to make my life easier and save me time/money. Use it for its intended purpose.

Depends on what you mean by normal but compared to many of the status pages I am tracking, GitHub has a pretty low incident rate (though an uptick very recently): https://statusgator.com/services/github

Nah, seems pretty normal. Compare to Netflix, Hulu, Heroku, etc, which have some outages nearly daily. This was just degraded performance for two hours, and the only more serious one I can recall was pretty recent.

dumb question but, why can't the teams of these super large companies who are probably getting paid top dollar get their deployment cycle right down to the point where they aren't causing interruptions?

Reply on news.ycombinator.com