This document summarizes the key findings from a survey on DevOps practices. It discusses the 5 stages of DevOps evolution that organizations progress through, from normalization of technology stacks to providing self-service capabilities. The document highlights that cross-team sharing is critical for scaling DevOps success, executives tend to have a rosier view than implementation teams, and organizations should start with automating processes close to production before addressing earlier stages. Automating security policies is also cited as important for achieving the highest levels of DevOps evolution. The survey gathered over 30,000 responses globally over 7 years.
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
The DevOps Playbook: How to Start, Scale, and Succeed
1. How to start, scale and succeed
Presented by Alanna Brown, Nigel Kersten, Andi Mann
and Michael Stahnke
2. About the authors
Alanna Brown
@alannapb
Nigel Kersten
@nigelkersten
Michael Stahnke
@stahnma
Andi Mann
@andimann
3.
4. 2018 State of DevOps Survey
● 3,000+ respondents
● Better global representation
● Translated in French, German,
Japanese and Malay
4
Survey responses by region
Over the past 7 years, we’ve gathered 30,000+ responses from around the world,
making the State of DevOps Report the largest and most comprehensive study on the topic of DevOps.
6. We want to hear from you!
● How did you expand DevOps success beyond a single team? What worked
and what didn’t?
● How do you collaborate with ITSM and ITIL in this new DevOps world?
● What’s hard with about automating incident response in your organization?
● How did you get executive buy-in? If you’re an executive, what do you need to
know to support your team as they adopt DevOps practices?
● How do you make sure your organization is doing monitoring and alerting the
right way?
7. Key findings
7
● In a DevOps evolution, there are many paths to success, but even more that
lead to failure.
● Executives have a rosier view of their DevOps progress than the teams they
manage.
● Start with the practices that are closest to production; then address processes
that happen earlier in the software delivery cycle.
● Cross-team sharing is the key to scaling DevOps success.
● Automating security policy configurations is mission-critical to reaching the
highest levels of DevOps evolution.
8. In a DevOps evolution,
there are many paths to
success, but even more
that lead to failure.
8
10. 5 Stages of DevOps Evolution
10
Stage 1: Normalization
Application
development teams
use version control
Teams deploy on a
standard set of
operating systems
Stage 2: Standardization
Teams deploy on a
single standard
operating system
Build on a standard set
of technology
Stage 3: Expansion
Individuals can do work
without manual
approval outside team
Deployment patterns for
building apps/services
are reused
* Infrastructure changes
are tested before
deploying to production
Stage 4: Automated
Infrastructure Delivery
System configurations
are automated
Provisioning is
automated
* System configs are in
version control
* Infrastructure teams
use version control
* Application configs
are in version control
* Security policy
configs are automated
Stage 5: Self-service
Incident responses are
automated
Resources available
via self-service
* Applications are
rearchitected based on
business needs
* Security teams are
involved in technology
design and
deployment
Stage 0
Monitoring and alerting are configurable by the team operating the service
Reuse deployment patterns for building applications or services
Reuse testing patterns for building applications or services
Teams contribute improvements to tooling provided by other teams
Configurations are managed by a configuration management tool
* These practices are highly correlated with the stage.
11. Stage 0: Build the Foundation
11
• Monitoring and alerting are configurable by the team operating the service
• Reuse deployment patterns for building applications or services
• Reuse testing patterns for building applications or services
• Teams contribute improvements to tooling provided by other teams
• Configurations are managed by a configuration management tool
12. Stage 0: Monitoring and alerting are configurable by team
operating the service
24xmore likely to always enable
configuration of monitoring and
alerting
13. Stage 0: Reuse deployment patterns for building applications
or services
23xmore likely to always reuse
deployment patterns for building
applications or services
14. Stage 0: Reuse testing patterns for building applications or
services
44xmore likely to always reuse
testing patterns for building
applications or services
15. Stage 0: Configurations managed by configuration
management tool
27xmore likely to always manage
configurations using a
configuration management tool
16. Stage 0: Contribute to other teams’ tooling
44xmore likely to always contribute
to other teams’ tooling
17. Stage 1: Normalize the Technology Stack
17
Stage Defining Practices Contributors to Success
● Application development
teams use version control
● Deploy on a standard set of
operating systems
● Build on a standard set of
technologies
● Put application configurations in
version control
● Test infrastructure changes before
deploying to production
● Source code is available to other
teams
1
18. Stage 2: Standardize and Reduce Variability
18
Stage Defining Practices Contributors to Success
● Deploy on a single standard
operating system
● Build on a standard set of
technologies.
● Reuse deployment patterns for
building applications and services
● Re-architect applications based on
business needs
● Put system configurations in version
control
2
19. Stage 3: Expand DevOps Practices
19
Stage Defining Practices Contributors to Success
● Individuals can do work
without manual approval
outside team
● Reuse deployment patterns
for building applications and
services
● Infrastructure changes are
tested before deploying to
production*
● Individuals accomplish changes without
significant wait times
● Post-incident reviews occur and results
are shared
● Build on a standard set of technologies
● Teams use continuous integration
● Infrastructure teams use version control
● Service changes can be made during
business hours
3
20. Stage 4: Automate Infrastructure Delivery
20
Stage Defining Practices Contributors to Success
● Automate system
configurations
● Automate provisioning
● Application configurations
are in version control*
● Automate security policy configurations
● Resources made available via self
service
4
21. Stage 5: Provide Self-Service Capabilities
21
Stage Defining Practices Contributors to Success
● Incident responses are
automated
● Resources available via self-
service
● Re-architect applications
based on business needs*
● Security teams are involved
in technology design and
deployment*
● Security policy configurations are
automated
● Application developers deploy testing
environments on their own
● Success metrics for projects are visible
● Experiences and lessons are shared
externally
● Provisioning is automated
5