Brainstorm general tool format

Initial thoughts on how the tool may work, be structured, etc.

changed milestone to %Sprint 100

added spike workflowSprint Ready labels

Thinking would be called with something like:

$ essm-tool --configuration production.yaml --really-do-it {sync action} {sync action} ... 

Configuration file (default to say essm-tool.yaml), would contain endpoints (testing or production), limits, etc plus a reference to where to get secrets.

{sync actions} would be: 'departments' - sync department from Booker to TermTime 'floors' - ditto for floors 'events' - scheduled events from TT to Booker 'block-bookings' - add/check block bookings in Booker etc

Caching of API calls so later actions can build on previous ones.

--really-do-it as opposed to an opposite --dry-run flag as by default you don't want to actually update but report what would have been updated?

Python files separating handling of termtime API usage, Booker API usage, configuration reading, common manipulation functions.

Retrieving secrets (and possibly configuration itself), could take principles from https://gitlab.developers.cam.ac.uk/uis/devops/django/externalsettings

Configuration could be a list of configuration files merged together, so secret and non-secret config can be stored separately.

If {sync action} could be a action that is really a list of other {sync action}s then the standard sync everything from B to TT, could be called, or just bits of it for testing/limited sync.

Before I over engineer this. Start with a tool that can read multiple configuration files and authenticate to APIs. Adding further 'actions' can be then split out. ==> #2 (closed)

closed

assigned to @rjg21