What does an Agile/DevOps organization look like?
I want to share an experience that highlights why I am so proud of the organization I work in. The experience is expressed in the form of an email thread that I had this week. Times/timezones are a little whacky because there were 3 – the customer’s, mine and the engineer’s. But the whole thread, from my perspective, started on Monday afternoon and ended by Tues night (and I was out of office on Monday afternoon or it likely would have been faster). I’ve written the thread backwards from most email clients so it is in chronological order and I’ve scrubbed all names, except mine, to protect the innocent 🙂
Sent: Monday, June 5, 2017 11:48 AM To: Brian Harry <email@example.com> Subject: VSTS web portal bug report from customer with a large number of dashboards Hi Brian, I wasn’t sure where to route this report, so I thought I’d bring it to your attention, from a user with a large number of dashboards. Any dashboard under the ellipsis menu (when there are too many dashboards for the screen width) isn’t currently selectable. The menu drops down, the URL in the browser window changes, but the actual view doesn’t so the user has to refresh the page. I’ve observed this happening with the user’s VSTS tenant in IE, Edge, and Chrome. I hope that’s clear and concise enough – let me know if I can do anything else to help. Thanks, —— From: Brian Harry Sent: Tuesday, June 6, 2017 9:38 AM Subject: RE: VSTS web portal bug report from customer with a large number of dashboards Thanks. XXXXXX, are you the right person to look at this? Brian —— Sent: Tuesday, June 6, 2017 9:40 AM Subject: RE: VSTS web portal bug report from customer with a large number of dashboards Yep, this is us. @XXXXXX can you please take a look. —— Sent: Tuesday, June 6, 2017 11:17:36 PM Subject: RE: VSTS web portal bug report from customer with a large number of dashboards The issue has been root caused and fixed. The fix will go out with the next release. —— Sent: Jun 6, 2017 4:35 PM Subject: Re: VSTS web portal bug report from customer with a large number of dashboards Wonderful, thank you for your quick work! I’ll let the customer know to keep an eye out in the next few weeks. —— Sent: 07 June 2017 01:04 Subject: Re: VSTS web portal bug report from customer with a large number of dashboards The release XXXXXX was referring to is a hotfix train that is typically scheduled each morning. This fix should be public by lunchtime tomorrow if all goes to plan. —— Sent: Wednesday, June 7, 2017 5:45 AM Subject: RE: VSTS web portal bug report from customer with a large number of dashboards Awesome, thanks – phenomenal turnaround! ————
Now, I don’t want to set the expectation that we can or do address every issue that quickly. But it’s not an accident that we were able to do this. It is many things:
1) An engineering team that is accessible to customers and wants very much to make them happy. 2) An efficient engineering flow that enables an issue to be identified, fixed and integrated into a “shipping branch” quickly. 3) An automated release process that enables improvements to be rolled out every day and reach customers running on infrastructure consisting of thousands of machines spread all over the globe. 4) A quality process that makes us feel confident enough that we can do that.
You often hear me talk about how we “release every 3 weeks”. That’s a great simplification of what we do. We have aligned 3 week sprints across our team and we ship “tip”, whatever has been checked in, roughly every 3 weeks. But, we have daily scheduled releases for medium priority fixes and, of course, on demand releases for urgent fixes. So literally, we ship more than once per day, on average. And we do that at huge scale. It’s taken us a lot of effort to get everything in place to be able to do that systemically and reliably at this scale but it’s a thing of beauty when it’s working right.
Just a thought that caught my fancy this week…