This document provides guidance on implementing DevOps changes incrementally. It recommends starting with small, low-impact changes that are easier to get approval for, such as focusing on a single bottleneck. The key steps are to pick one thing to work on, assemble a small champion team, define success metrics, research solutions from other organizations, test potential tools, make a selection based on proving the concept works, and then roll it out as a standard practice before identifying the next improvement. It acknowledges that failures may occur, and advises trying a different approach rather than giving up, as DevOps is a continuous journey.
2. BIG THINGS HAVE SMALL BEGINNINGS
- T. E. LAWRENCE IN LAWRENCE OF
ARABIA
3. TEXT
WHY IS INCREMENTAL OFTEN BETTER?
▸ Easier for management buy-in
▸ Easier to be under the radar — less visible to those who are
resistant to change
▸ Fewer approval / budget constraints
▸ DevOps is a journey, not a destination
4. TEXT
WHEN IS INCREMENTAL NOT THE RIGHT APPROACH?
▸ When things are so bad you don’t have time to get better
over time
▸ When you have a mandate to deliver big results under a
deadline
▸ Sometimes, there’s no way to dip your toe into the water —
the only way is to accomplish your goal is to jump in
6. HOW TO INCREMENT YOUR DEVOPS
PICK **ONE** THING TO WORK ON
▸ Don’t try to do too much at once. You will never do anything if
you try to do everything
▸ Analyze your bottlenecks. Where is work slowing down? Are
projects always waiting on one person, or on one team?
▸ Think about areas where changes could be low-impact, and
possibly visible to other parts of the business. Think about if
there are multiple groups that could benefit from the area of
focus
▸ Don’t pick a tool and then try to find a problem to solve
7. HOW TO INCREMENT YOUR DEVOPS
PICK THE TEAM
▸ Not too big, not too small
▸ You need one person who can make decisions
▸ You will be better off if you can find a champion in
management
▸ You need a facilitator or PM to keep things going
▸ Don’t necessarily exclude the troublemaker, but there is a
fine line between critical questioning and sabotaging
8. HOW TO INCREMENT YOUR DEVOPS
TREAT THIS LIKE A REAL PROJECT
▸ Define the plan: who will be involved? What timeframe will
you work from? How much time can the team devote to it
per week?
▸ Often internal and POC projects are low priority, but the
more you can treat it like a real project, the better the
chance of success
9. HOW TO INCREMENT YOUR DEVOPS
WRITE THE ENDING FIRST
▸ If you don’t define what success looks like before you begin, it’s very
easy to keep redefining your endpoint as you go along
▸ Define short-term goals and long-term goals
▸ Think about why you haven’t been doing this already. Are there
roadblocks to avoid? Who is threatened the most by this change?
How can you either demonstrate benefits to those people, or avoid
them until you can demonstrate success to others?
▸ This is incremental, so the ending shouldn’t be “and they all lived
happily ever after”. It should be more like “deployments take 15
minutes instead of 2 hours”.
10. HOW TO INCREMENT YOUR DEVOPS
RESEARCH WITH CONTEXT
▸ There are few unique problems that other organizations haven’t
encountered and solved. Are you trying to reduce deployment
issues? Look at companies who have solved this well.
▸ BUT — you are a special snowflake, so don’t bother with possible
solutions that would never fit. If you work in a conservative company,
that might mean open source solutions won’t be well received.
▸ Don’t research too long. There are new tools every day. Newer isn’t
better. Timebox your research, and make a decision
▸ Don’t make the final decision during research — pick a few options to
test
11. HOW TO INCREMENT YOUR DEVOPS
PROOF OF CONCEPT
▸ Take test drives of your contenders in your environment
▸ Determine how to objectively measure which one solves the
problem best. Decisions shouldn’t be swayed by a shiny UI
or popularity / coolness
▸ Test test test
▸ consider scenarios where the solution will not work. How
will those situations be handled?
▸ Measure against your definition of success
12. HOW TO INCREMENT YOUR DEVOPS
DECISION TIME
▸ You have a winner! Based on your careful research and
testing, the team found a solution to make your
organization better/faster/stronger
▸ Your work isn’t done:
▸ how are you going to make this a standard practice or
tool for your team?
▸ how are you going to sell it to your team? To other
teams?
13. HOW TO INCREMENT YOUR DEVOPS
ROLL OUT AND DO IT AGAIN
▸ If everything goes well, hooray!
▸ This is incremental (remember?), so repeat from the top
15. TEXT
FAILURE IS ALWAYS AN OPTION
▸ Projects never go according to plan.
▸ (now is when we vent)
16. TEXT
THE SUN WILL COME OUT TOMORROW
▸ Back to incremental — did an initiative fail? Don’t give up,
just go smaller or in a different direction.
▸ Think back to a previous success, and what allowed it to
succeed. Can you try again using a different approach
17. TEXT
RESOURCES / INFLUENCES ON FOR THESE SLIDES
▸ “Incremental Devops” http://inedo.com/incremental
▸ Sprint: Solve Big Problems and Test New Ideas in just FIVE
DAYS http://www.thesprintbook.com/