avatarWillem-Jan Ageling

Free AI web copilot to create summaries, insights and extended knowledge, download it at here

2149

Abstract

end-to-end testing. Added to that we focused on creating smaller stories that we could get to “Done” in a sprint.</p><p id="e212">So we knew our intended direction. Below I describe the steps in our journey to get there.</p><p id="b3a8"><b>Step 1</b></p><p id="a813">We started with the following:</p><ul><li>Reduce the size of the backlog items via story slicing.</li><li>Arrange close cooperation with UAT team</li></ul><p id="735d">These were low impacting changes for the organization, but we reduced the average time to production with 3 days.</p><p id="ee06">It did not extend our Definition of Done though. We still had the same skills NOT under control as we had at the start.</p><p id="2e71"><b>Step 2</b></p><p id="cfa7">We then focused on:</p><ul><li>Bringing UAT within the team</li><li>Automated deployment</li></ul><figure id="ba9e"><img src="https://cdn-images-1.readmedium.com/v2/resize:fit:800/1*fNkjb-Oi_hp8NHwHio7XIg.jpeg"><figcaption></figcaption></figure><p id="553f">Here we had our first change that impacted the organization. UAT moved from the business department to the Scrum teams.</p><p id="f1c4">The automation of the deployment required consireable effort from the team. But the investment soon showed to be worthwile.</p><p id="d11a">We reduced the average time to get an item to production with another 2 days. So in totaal we went from 3 weeks to 2 weeks. The average duration to get an item to production now was within the Sprint, but because we didn’t have the full chain into the Scrum team’s control we still could not expand the Definition of Done.</p><p id="f780"><b>Step 3</b></p><p id="e7f7">This is where we are now:</p><ul><li>Bring deployment to pre-production and to production environments into the team.</li><li>Automated end-to-end testing.</li></ul><figure id="d45d"><img src="https://cdn-images-1.readmedium.com/v2/resize:fit:800/1*8yKyGeSJWWMYZfHwYtmVMg.jpeg"><figcaption></figcaption></figure><p id="1e90">This involves the second organizational change. People will move from Operations to the Scrum teams.</p><p id="76e0">We will gain 2 to 3 days with the automated testing and 2 days with the

Options

deployment capacity within the teams.</p><p id="b48b">When we are finished with this step we will be able to finally expand the Definition of Done all the way up to production.</p><p id="387b"><b>Step 4</b></p><p id="9ad8">With step 3 we get the whole process from idea up and until deployment to production within the team. To have a true Devops environment we will include:</p><ul><li>Incident resolution responsibility will be transferred to the Scrum teams.</li><li>Other Operations capacities like networking, database administration etc should be incorporated in the Scrum teams.</li></ul><p id="c82e">We are in quite a journey towards Devops by extending our Definition of Done. This post may make it seem easy but there’s a lot of negotiation, coordination and work required.</p><p id="53fe">All our teams are different, impacting the speed of transition. I am involved in three teams. One has the advantage that items can be sliced to be very small. Their journey has been easier than I described. Another team has to deal with many more interdependencies. They will need more time. The third team, the team of the example, did have many opportunities to improve relatively rapidly. The most interesting story.</p><p id="388b"><i>I believe that Scrum and Devops go hand in hand.</i></p><p id="302f">It should be the ultimate goal to expand the Definition of Done all the way to production. Including tasks that are at first a responsibility of Operations. Your end result will be a Devops Scrum team!</p><p id="d3df"><b>Did you like the article? Then it would be awesome if you’d clap 👏🏻. I am also very keen to learn what you think about this topic.</b></p><p id="348a">My twitter profile is <a href="https://twitter.com/WJAgeling">https://twitter.com/WJAgelin</a>g</p><figure id="a703"><img src="https://cdn-images-1.readmedium.com/v2/resize:fit:800/1*qsg-zjcnz5A8B1xmBbdIfw.png"><figcaption><a href="https://readmedium.com/your-invitation-to-the-serious-scrum-slack-workspace-f424aeea4093?sk=e8334e6ee505a85ae6b9d2a1ce37219c">Do you want to write for Serious Scrum or seriously discuss Scrum?</a></figcaption></figure></article></body>

Scrum and Devops going hand in hand — our journey

Devops and Scrum – extending the Definition of Done brings us Devops

We do Scrum and… — episode 4

The Definition of Done (DoD) should be a defined concept within a team. Everyone in the team should know what is included in the DoD. But also what is excluded.

At the moment that our Scrum team started we had developers and testers within the team. So our definition of Done included everything that could be controlled by development and testers.

When our potentially releasable product was “done” we still needed to:

  • Deploy on pre-production
  • Perform User Acceptance Test
  • Deploy to production

We also had other factors that impacted what we could do within a Sprint:

  • No automated deployment of the code
  • No automated end-to-end testing
  • Big backlog items/user stories

The result of the above was that we spent quite some time on manual work and handovers to other departments. It took us approximately the time of a half a Sprint, 1 week, to get the code from “Done” to production.

We believe that the whole idea of Scrum is that the development team has all the skills necessary to create a production increment.

In our case this includes UAT and deployment capability to pre-production and to production. This required an organizational change because both capabilities were at different departments. UAT resided within a business department and deployment within Operations.

But having all the skills within the development team to bring an item to production is one.

To bring an item from idea to production within a sprint is the other aspect of the Definition of Done.

For us it meant that we had to work on automated deployment and automated end-to-end testing. Added to that we focused on creating smaller stories that we could get to “Done” in a sprint.

So we knew our intended direction. Below I describe the steps in our journey to get there.

Step 1

We started with the following:

  • Reduce the size of the backlog items via story slicing.
  • Arrange close cooperation with UAT team

These were low impacting changes for the organization, but we reduced the average time to production with 3 days.

It did not extend our Definition of Done though. We still had the same skills NOT under control as we had at the start.

Step 2

We then focused on:

  • Bringing UAT within the team
  • Automated deployment

Here we had our first change that impacted the organization. UAT moved from the business department to the Scrum teams.

The automation of the deployment required consireable effort from the team. But the investment soon showed to be worthwile.

We reduced the average time to get an item to production with another 2 days. So in totaal we went from 3 weeks to 2 weeks. The average duration to get an item to production now was within the Sprint, but because we didn’t have the full chain into the Scrum team’s control we still could not expand the Definition of Done.

Step 3

This is where we are now:

  • Bring deployment to pre-production and to production environments into the team.
  • Automated end-to-end testing.

This involves the second organizational change. People will move from Operations to the Scrum teams.

We will gain 2 to 3 days with the automated testing and 2 days with the deployment capacity within the teams.

When we are finished with this step we will be able to finally expand the Definition of Done all the way up to production.

Step 4

With step 3 we get the whole process from idea up and until deployment to production within the team. To have a true Devops environment we will include:

  • Incident resolution responsibility will be transferred to the Scrum teams.
  • Other Operations capacities like networking, database administration etc should be incorporated in the Scrum teams.

We are in quite a journey towards Devops by extending our Definition of Done. This post may make it seem easy but there’s a lot of negotiation, coordination and work required.

All our teams are different, impacting the speed of transition. I am involved in three teams. One has the advantage that items can be sliced to be very small. Their journey has been easier than I described. Another team has to deal with many more interdependencies. They will need more time. The third team, the team of the example, did have many opportunities to improve relatively rapidly. The most interesting story.

I believe that Scrum and Devops go hand in hand.

It should be the ultimate goal to expand the Definition of Done all the way to production. Including tasks that are at first a responsibility of Operations. Your end result will be a Devops Scrum team!

Did you like the article? Then it would be awesome if you’d clap 👏🏻. I am also very keen to learn what you think about this topic.

My twitter profile is https://twitter.com/WJAgeling

Do you want to write for Serious Scrum or seriously discuss Scrum?
Scrum
Agile
DevOps
Serious Scrum
Scrumand
Recommended from ReadMedium