Continuous Delivery and Quickstart guide Highlights
Delivery
Software Delivery Process Flow 1.0 and 2.0 do not have monitoring defined
Monitoring generally not added until SDF 3.0 way too late
Change
Never Break your consumer
Continuous Integration
Is a must for Continuous Delivery
Environments
The ideal number of Environments is two, development and production
CD Tooling
CD Tool features list see book
No Downtime Deployments
Deployment downtime is unforgiveable
Maximum Uptime
Maximum uptime is required for all environments, development and production
Monitoring
Monitoring is a must for CD and DevOps to work correctly
Virtualization
If virtualization supports your platform / application then use it
Departmental Meetings
If problems are discussed in dept. meetings instead of 5 min. desk discussions cancel then
Building Trust between Org Boundaries hard to do
Central change control bag, decent change notification good, change notification is not change control
Trust is earned over time by peoples actions
If we are consistently told to work long hours and weekends there is something wrong
With the priority of the work
The Majority will be behind you but we are doing something we have already asked for
Change impacts people in different ways and some react with hostility
Outsiders to the project should be involved as well, Marketing for example
Relentless patience is not an oxymoron
CD and DevOps requires change , but the change is good, for people and systems
Be highly transparent and visible to what changes will be happening inside and outside the CD / DevOps project.
Potential Issues to look out for (see book)
Hostile staff, Dissenters, Red Tape, Geographical diversity
Use face to face discussions to overcome, not mail
Use PR team to writearticles to improve project visibility
Project interview question – see book how would you feel if your code was running prod. 30 min .after in source control
Monitoring should as early on in Project as possible, will become an integral part of CD DevOps
Development and Production environments should be monitored the same way
Combine automated tests and system monitoring, infrastructure monitoring .
System Monitoring and Infrastructure Monitoring should be deployed side by side
Monitoring results should be complied into Reports andGraphs
Monitoring Metrics – see book
Number of deployments per period
Release time
Commit to Production time
List of unique software components
Last Step Train more CD DevOPs Trainers