Skip to content

Automatically deploy each main detector build to production and recovery - #18

Merged
matteius merged 1 commit into
mainfrom
fix/automatic-hosting-deployment
Oct 30, 2026
Merged

matteius merged 1 commit into
mainfrom
fix/automatic-hosting-deployment

Conversation

@matteius

Copy link
Copy Markdown
Contributor

Every main push currently publishes a detector image but leaves both production origins on the previous build. This adds an automatic paired release agent and makes CI wait for its exact-build deployment receipt.

The agent warms cloud before handing over traffic, updates both origins and the recovery controller to the build's immutable amd64 digest, returns traffic locally, and verifies the persistent cloud model cache before scaling cloud back to zero. Durable checkpoints resume interrupted handoffs; another operator's hold is never overwritten. The operator-host timer is installed and follows main, so merging activates the path without another deployment approval.

Validation: six regression tests cover artifact/source identity, stale-image rejection, terminating pods, interrupted handoffs, operator ownership, and platform manifests. Python compilation, shell syntax, and git diff checks pass. Forms and detector application credentials remain unchanged.

@matteius
matteius merged commit bd80971 into main Oct 30, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant