Comment by Someone1234 | Hacker News Reader

Comment by Someone1234 | Hacker News Reader

I wonder if Azure DevOps will have a day long outage like VSTS had last week? That had a legitimately productivity cost to a lot of companies, and the more vested in the VSTS toolchain you were the harder hit you were.

I'm legitimately surprised they're launching this so soon after that, particularly considering there's been no real post-mortem, and no update on how they'll stop this happening again.

"It is Azure's fault" isn't an answer, and re-branding this Azure DevOps makes that excuse almost hilarious. I'm well aware that Azure AD was down, and that VSTS has a strong dependency on it, but ultimately we're still talking about a single data center bringing down VSTS for all of North America for an entire work-day.

It has been a week and now they're trying to encourage you to push your whole deployment toolchain (manage, build, test) into their cloud.

I certainly hope not. I'm not going to argue that this was anything but a disappointment to our customers - and one that we're eager to prevent from happening again.

But we have never said that this was Azure's fault. It wasn't. We've put the blame where it belongs: on us. We're nearly done writing up the root cause analysis and an analysis of our next steps to prevent this from happening in the future.

Do you happen to know where the root cause analysis will be posted if it is made public?

It will be on our service availability blog: https://blogs.msdn.microsoft.com/vsoservice/, and I'll reply here with a direct link once it's online.

Update: the Director of Engineering for Azure DevOps has posted this. https://blogs.msdn.microsoft.com/vsoservice/?p=17485

3 weeks ago there was a different outage that meant no hosted build agents were working in the west europe region, and it lasted for more than 24 hours.

Now, what I hated the most about this was the communication - there wasn't any. Why not email impacted users to let them know you're aware and working on a fix?

You can subscribe to get emails for availability incidents on our service availability blog, https://blogs.msdn.microsoft.com/vsoservice/ (it's in the right-hand sidebar).

Huh, I hadn't noticed that - I've subscribed now.

I do still think it would be better if you proactively sent out emails only to affected customers though.

Reply on news.ycombinator.com