Deploy UGRC API to production

[steveoh]

Benefit

A cloud based system using modern supported technology with improved documentation and self service websites

Acceptance Criteria

• production milestone issues
• The DNS is swapped to the cloud system
• the on prem system is turned off

Notes

No response

Risks

The new API could break existing clients

Issue Reference

refs #309

[steveoh]

Sprint 1 Notes (Steve)

July 3

• All systems have been deployed and configured.
  • The forklift pallet needs an update to notify the prod system to update.
  • The agrc shared account is registered and has claimed old account keys to verify that is working.
• The notice email is scheduled for the 8th.
• I stress tested the API to a point that I think is acceptable.
  • We can always make the server bigger if need be. It's CPU constrained.
  • Future improvements to caching will help alleviate pressure on this machine.
  • This will require monitoring after deployment.
• I dumped the current user store into the production system and will need to refresh it prior to the DNS switch.
• I did not load the redis stats but will do so prior to the DNS switch.
• I deleted the memory store and stopped the compute engine in dev to save on cost.

[steveoh]

Sprint 1 Notes (Steve)

July 15

• Notice email was sent
• Change management scheduled for an august 6th deployment

[steveoh]

Sprint 2 Notes

July 29

• Existing database migration code updated to run from the source database and fill in any missing data
• Usage stats needs to be loaded from a dump on an authorized cloud network and will be done just after 10am on the 6th
• Change request tasks (CHG0039644) continues to evolve and progress
• Service now task is entered to schedule DNS migration - SCTASK0026297
  • Firebase hosting verification process has started but not completed
  • TTL for DNS will be updated to 5 minutes ~24 hours prior to DNS switch to ensure a speedy update
• On-prem system will be powered off on the 8th
• Decided to send an email on the 6th in the afternoon to notify users of the change and requirement to update their authentication and claim their keys.
• Need to figure out who manages azure devops to remove CI pipeline

[steveoh]

Sprint 3 Notes

• The API was deployed to production with the new self service and documentation sites
• Emails were sent to notify users to claim their legacy account data
  • People had trouble remembering passwords and we migrated keys for trusted users
• Bug fixes were deployed as they were identified
• Website analytics were added to the websites
• A dashboard was created to view api analytics
• Legacy cloud resources (functions/recaptcha) and CI system was shut down. SCTASK0026466, SCTASK0027660, SCTASK0027450
• The api client was updated to reflect the latest changes
• We're leaving the on prem api on a bit longer to diagnose discrepancies in api features/responses.
• VIP diagrams updated

[steveoh]

Sprint 4 Notes

• shut down on prem machine RITM0275976