How to play

Download the rules in the resources section on the right sidebar

Setup

Space: You will need a large space with 2 tables to play.

As a facilitator explains the game from the rules, then:

• Separate physically the Dev and Ops teams
• Put the “Dev” environment on the dev table
• Put the “Pre-production” and “Production” environments on the ops table
• Add the base structure for tower bases on each environment. The base must be as in the image below : (base structure represents the environments (DB, Frameworks, Languages, …)) 

Rules

Sprints: There will be 4 sprints during this game. Each sprint will be organized as described in the sprint image. For each sprint, the facilitator must distribute the corresponding cards (# sprint number).

Goal

The goal is to aim for maximum points by delivering features to production (deployment). We will use wooden bricks to do so.

Roles

Sprints

Sprint 1: Documentation approach vs collaboration

• DEV: 3 minutes to build and deliver to pre-prod
• OPS: Refuse any non-documented tower

Retrospective

• Stop starting, start finishing / Work In Progress (W.I.P) limits
• Keep It Simple & Stupid (K.I.S.S)
• Production deployment requires collaboration and reveals problems.
• Silo breaks the collaboration

Sprint 2: Silo again

No collaboration: it is forbidden to communicate with Ops (by the top management)

• DEV: Document the delivery + no cards for them
• OPS: T shape base in pre-prod + prod

Retrospective

• Opposite objectives between Dev & Ops (accounting, different priorities/projects/visions)
• Definition of done
• Always think about the targeted environment

Action

Delivery and deployment in the presence of both teams.

Sprint 3: Culture of collaboration

Move from siloed delivery to collaboration: everyone in one room.

• DEV: construct, starting by taking back the previous tower + cards for Sprint 3
• OPS: facilitates the deployment in pre-prod + prod

Retrospective

• Collaboration makes it possible to deliver
• Focus on culture/collaboration
• Collaboration saves time

Sprint 4: Automation

• DEV: construct, starting by taking back the previous tower + cards for Sprint 4
• OPS: automate the deployment between pre-prod and prod

Retrospective

• Clone of production (blue/green deployment), could be simulated by swapping two pre-production and production post-its
• Automate deployment and acceptance testing, instead of manually doing it
• Pre-production environment for practicing before production
• Continuous deployment card by card, the ops manager sums the deployment times
• Automation saves time

Conclusion

• Evolution of the metrics used on the boards, and aggregation into one shared board (performance, process, people, KPI linked to events)
• Collaboration is a skill that can be learned
• Explain the basic principles of DevOps (Be C.A.L.M.S)