The 7 Principles of DevOps and Cloud Applications

THE 7 PRINCIPLES OF
DEVOPS AND CLOUD
APPLICATIONS
Gerardo A. Dada
VP, Product Marketing & Strategy
SolarWinds

What is DevOps?
What can IT learn from
DevOps?
How can I apply DevOps
thinking to my current IT
environment?
2
W H AT TH I S P R E S E N TATI O N AI M S TO AN S W E R
© 2015 SOLARWINDS WORLDWIDE, LLC. ALL RIGHTS RESERVED.

WHAT IS DEVOPS?
There is no official definition of DevOps, there are many.
DevOps allows to respond faster to customers, fix things faster,
produce with more quality. Quality includes performance, security and
less bugs.
However, you don’t ‘do’ DevOps, it’s not a process, it is a business
practice, an approach.
DevOps cannot be stealth, cannot be grassroots. It requires support
from top to bottom, it changes the entire system, This is why DevOps
starts with people and process first. If you don’t have culture, all
automation attempts will be fruitless.
3© 2015 SOLARWINDS WORLDWIDE, LLC. ALL RIGHTS RESERVED.
AN D W H AT C AN W E L E AR N F R O M I T?

DEFINING DEVOPS
“DevOps is a software development practice where development and
operation teams work together, taking the intelligence of how an
application runs to inform and improve how the application is being
built, in a rapid iterative process.
It embraces the concepts of continuous delivery and continuous
integration and requires a strong discipline around performance
optimization, monitoring and automation.”
Let’s break this definition down into 7 principles

PRINCIPLE #1 END USER FOCUS
DevOps is a software development practice where development and
Application and End-User Focus - Everyone on the team is focused on
making the application run well and its impact to end.
• Requires thinking about a system’s end-to-end processes, considering
all the steps a software change must take, from the client’s initial
request until the deployment to production. This helps to avoid local
optimizations and to bring down work silos
• Infrastructure is ephemeral, dynamic, and utilitarian. It serves only to
make the application work.
• The metrics that matter are application metrics and end user
experience.
• Application metrics are connected to key business metrics. Software
becomes a driver of the business.
I T‘ S AL L AB O U T TH E AP P AN D E N D U S E R E X P E R I E N C E

PRINCIPLE #2 COLLABORATION
application runs to inform and improve how the application is being built, in
a rapid iterative process.
Collaboration – Because goal is the end user, silos do not work.
• If the application is down everyone has failed.
• There is no virtualization department and storage engineer. There is
only one team: the one responsible for the app to work.
• Not only are there no silos, teams must work together to understand
how the system works as a whole and how each element impacts
performance
• This requires transparency, visibility, a consistent set of tools and
teamwork. And it requires expertise across the stack.
D E V E L O P, TE S T AN D R U N - AN I N TE G R ATE D P R O C E S S

PRINCIPLE #3 PERFORMANCE ORIENTATION
application runs to inform and improve how the application is being built,
in a rapid iterative process.
Performance Orientation – Performance is a requirement and a
discipline.
• Uptime is not in question, is assumed.
• Performance is measured, all the time, everywhere.
• Everyone on the team understands performance, and the implications
of bad performance, from end-user experience to business impact.
• There is a shared understanding of what performance metrics are
important and what are the performance thresholds and goals.
• Bottlenecks and contentions are well understood.
• The team understands there is a correlation between performance and
cost.
• Performance is managed and planned for, as the team plans for
resources and growth (usage and capacity).
P E R F O R M AN C E AS A D I S C I P L I N E

PRINCIPLE #4 DEVELOPMENT SPEED
DevOps is a software development practice where development and operation
teams work together, taking the intelligence of how an application runs to inform
and improve how the application is being built, in a rapid iterative process.
Speed –Accelerate time to market of new software. Deploy new software every day,
multiple times.
• Short, iterative processes allow teams to move faster. Break big problems into
smaller chunks.
• Accelerate innovation and learning. Serve the business more effectively. Give
people what they ant, faster. From idea to production in hours.
• Software that is not in production is useless. It is like inventory sitting in a
warehouse getting old.
• Every development sprint produces quality code. Every change and every
release is tested, measured, pushed to production and measured all the time.
The impact of changes is understood. Quality is an ongoing process.
B U I L D FAS TE R . FAI L FAS T. L E AR N Q U I C K .

PRINCIPLE #5 SERVICE ORIENTATION
Service orientation - Everything is a service: from application components
to infrastructure.
• Everything is programmable. Software controls everything.
Infrastructure as code.
• Servers are no longer pets you care for and fix, they are cattle. If they
get sick, you shoot it and get a new one.
• Infrastructure and new capabilities are one API call away.
• Everything is flexible, can change quickly and is ready to scale.
• No monolithic apps. Components are asynchronous. Systems are
designed for horizontal (and vertical) scalability.
• Everything can fail, resiliency is built in. When something fails, it is
replaced by a new instance that takes the place of the failed one.
• Things are broken on purpose to test for resiliency.
L O O S E LY C O U P L E D S E RV I C E O R I E N TE D C O M P O N E N TS

PRINCIPLE #6 AUTOMATION & REPETITION
application runs to inform and improve how the application is being built, in
a rapid iterative process.
Automation – Automate everything that can be automated.
• There is a culture of continuous experimentation.
• Repetition is the only way to achieve mastery. Architect things to be
repeatable.
• Practice, practice, practice: from code deployments to scaling and
failovers,
• Automation enables speed and consistency and allows the team to
focus on what matters.
• Infrastructure is controlled by software. Code, deployments, tests,
monitoring, alerts – everything is automated.
• Continuous development and continuous integration aided by
configuration management tools like Chef, Puppet, Salt and Ansible.
• Automate IT services for end-users: End user self-service. IT as a
Service.
AU TO M AT E E V E RY TH I N G TH AT C AN B E AU TO M AT E D

PRINCIPLE #7 MONITOR EVERYTHING
Monitor everything - Monitoring is a requirement and a discipline.
• Monitoring is the enabler. Provides intelligence and gets all members of
the team on the same page.
• DevOps dictates an amplification of feedback loops so that any
problems are quickly identified and corrected
• If you can’t measure, you can’t improve
• Visibility is critical for speed and collaboration.
• Everything is tested. The impact of every change in known.
• Dashboards show metrics that matter, not just what can be measured
• “If it moves, measure it” …“If it’s not monitored, it doesn’t exist”
M O N I TO R I N G I S AN E N AB L E R AN D A D I S C I P L I N E

7 DEVOPS PRINCIPLE
1. Application and End-User Focus
2. Collaboration
3. Performance Orientation
4. Development Speed
5. Service orientation
6. Automation
7. Monitor everything
12
D E V O P S P R I N C I P L E S Y O U C AN E M B R AC E TO D AY
© 2015 SOLARWINDS WORLDWIDE, LLC. ALL RIGHTS RESERVED.

THE ROLE OF OPERATIONS TEAMS
How are things working? How do I improve performance?
How do I fix problems fast?
» Gather telemetry or metrics on the system/app
» Use data to make decisions on capacity, infrastructure, & architecture
» Know what the bottlenecks and dependencies are
» Know when thresholds will be hit
» Advice the business on cost/performance tradeoffs
» Fix things when they break - quickly (incident management)
 Identify the problem and its severity
 Find the root cause, what caused the problem, in what part of the stack
 Fix it
13

THE PERFECT DASHBOARD
14
How fast we are
going?
page load time
How hard is the
engine working?
throughput, CPU
utilization
Anything
broken?
alerts
How fast can
we go?
redline
How long can
we go?
threshold
Credit: David Lutz via slide share

• Turning data to insight – what does this mean?
• The orange engine light is on, what does it mean? Do I need to stop?
• False alarms and alerts – relevance and criticality
• The annoying ‘ding, ding, ding, ding…” when the car door is open
• Identifying causality – root cause
• The car is sputtering. Is it the gas? A clogged fuel injector? Bad spark plugs?
• Different tools, different opinions – finger pointing
• The tire pressure light is on, but my pressure gauge says it’s fine
• Complexity and scale
• Managing hundreds or thousands or rapidly changing resources
• Managing growth and change
• Most monitoring systems are a a Frankenstein of duct-taped components
15
MONITORING CHALLENGES
C O M P L AI N T S AB O U T M O N I TO R I N G

The 7 Principles of DevOps and Cloud Applications

More Related Content

Viewers also liked (20)

Similar to The 7 Principles of DevOps and Cloud Applications (20)

More from SolarWinds (20)

Recently uploaded (20)

The 7 Principles of DevOps and Cloud Applications