Octopus energy has 500 developers
5000 employees in total [biggest supplier of energy in UK]
valued at £8bn after 8 years

160 releases each day
2500 deployments a day
— heavily automated and automatic tests

simplicity over everything
dont do something more complicated till much later on
minimum viable EVERYTHING
high trust, but with audit trail

developers are responsible for the whole cycle (and deploying it), and code reviews (deployments are on merge requests) and devs are responsible for verifying deploy is complete and changes are correct.

they have no QA dept (but devs do QA)
QA via pairing, automated tests, static analysis, code review

no meetings for developers – keep them on their work (same with the general staff)

they track technical debt as a KPI
they do one day a month, all devs stop feature-release type work, and they do a ‘spa’ day, where they just fix as mangy unresolved bugs as they can in a day.

they use the fixing-bugs thing as a way to onboard new users
how to triage
how to debug
how to resolve
how to deploy
how to get merge requests etc

** merge request means commit to main
** that means deploy

only put things in sentry that they can fix
they use a lot of linting and auto things
don’t allow a 500 error to be induced (they work on this a lot)

alert-to-resolution
they should have great pull request messages
they have a syntax for sentry to mark the issue as resolved automatically from git messages (so sentry is listening to their git requests) – should that be a different project, or the same one – same one probably

tech-debt-burndown
how do they prioritise (the staff decide)
– by frequency
– by customer impact
– by embarrassment factor

the goal is to get to zero eventually, but there are hundreds of thousands of ongoing errors and issues (some errors happen more than a million times before they address them) – they have billions of errors!