Juju Team Updates - Week #14 2022

Hi everyone, below you can find the updates from the Canonical Juju Team for the week starting on the 04/04/2022. Have a great weekend!

cc @pedroleaoc @jameinel

  • explored the possibility of using the Snap Profile/aspect approach to design Juju service locators
  • investigated AWS and GCP networking models to provide the model abstractions for the network activity in Juju
  • a bit of bugs triage and code review


  • More clean-up work in the Juju OLM Reference and How-to guide docs.
  • Initiated plans for similar overhaul of Juju SDK Reference and How-to guide docs. The idea is for the Reference docs to document all the machinery (Ops, Charmcraft CLI commands, Concepts) and then to have cross-referencing between these docs and the how-to guides that rely on them.


  • Progress with the tutorial.

Landed all but one of my PRs before signing off from the Juju team.

  • Landed the dqlite development into a new 3.0/develop branch
  • Updated the logging to allow forwarding to syslog instead of the database to see if logging is actually the bottle neck in production.
  • Fix LXD errors for already existing containers
  • Remove reflection when creating a facade, this should improve performance and when starting a controller.
  • Allow proxy aware charmhub clients
  • Allow centos to be fully integrated when deploying from charmhub

See: https://github.com/juju/juju/pulls?q=is%3Apr+author%3ASimonRichardson+is%3Aclosed

So Long, and Thanks for All the Fish

  • 2.9.28 release
  • many changes to our build / release jenkins and associated repos so that
    • release jobs now all run on ephemeral nodes
    • port various scripts to Python 3
    • more precise selection of various credentials and prerequisite packages
    • release process has a dry-run option for testing
    • we are ready to be ale to move the release process to a new jenkins when the data centre hosting the current servers shuts down
    • lots of cleanup of legacy code / stale credentials / obsolete scripts etc
  • Patch to ensure that changes in cloud availability zones are accommodated by Juju’s provisioner each time it is woken up to provision machines.
  • Kicking off work on a relational schema as part of the transition away from MongoDB.
  • Other debrief and end-tying related to @simonrichardson’s last week. :wave: :cry: