Devops

From Federal Burro of Information
Revision as of 20:30, 16 July 2021 by David (talk | contribs) (→‎blogs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

the twelve factor app : http://12factor.net/

https://try.newrelic.com/rs/newrelic/images/The_New_Kingmakers.pdf


DevopsDays 2015 Conference Notes


blogs

https://charity.wtf/2019/10/28/deploys-its-not-actually-about-fridays/

http://michal.karzynski.pl/

https://www.arresteddevops.com/

https://www.adaptivecapacitylabs.com/blog/2020/12/21/incident-phenomena-shorthand-names-a-la-danny-ocean/

Continuous Integration

  • Is everyone committing to the same trunk at least once a day?
  • Is every commit being tested?
  • When the build breaks Is is red fixed in ten minutes?

Running Jenkins against you feature branch and then ignoring the build when it turns red is not Continuous Integration

Continuous Deployment

when the CI build passed it goes to prod automagically

  • increate dicipline
  • increate hygene
  • reduce time to market
  • decreese feedback loops
  • Improve quatly
  • Reduce Cost

You should do this if you can, sometime you can't: firemware , mobile apps

Continuous Delivery

  • for firmware


Patterns

Canary
Love this , for software roll outs.
must be-able to have concurrent versions of app deployed at the same time. Not always possible. Test before hand? ha wouldn't that be nice.
blue green
Love this, but for infrastructure roll-outs not software package updates.
For example I might do this for a major kubernetes version update.
strangler pattern - for moving to
circuit breaker
https://martinfowler.com/bliki/CircuitBreaker.html