Every deploy of one service fails instantly at CREATE_CONTAINER ("Failed to create deployment."), no logs

Corpo (colar como está):

Since 2026-07-20 ~13:28 UTC, 8 consecutive deployments of ONE specific service fail instantly (0s) at "Deploy > Create container" with only "Failed to create deployment." - no build logs, no deploy logs.

Project: destrava-agent (7bdcab83-3b15-47b6-92cf-da8f58688899)

Environment: production (f33f39cd-59ff-4d83-9a69-0ac457a48598)

Service: evolution-api (9f8fffa9-07ba-4434-a2e8-c8ab0015026f) - Docker image service (atendai/evolution-api)

Failed deployment IDs: 17d77c62, 9b29526f, f0a7fcef, 563346c3, c3e5c039, b5c98e25, b5220246, 6e61346e

Ruled out on our side: env var change vs redeploy vs image pin (latest -> v2.2.3) vs region change (us-west2 -> us-east4) - all fail identically; env values validated; control tests in the SAME project today: repo-build service (Worker) redeploys fine AND a brand-new Docker Hub image service (nginx:alpine) deploys fine. In-dashboard AI diagnosis says "platform-side scheduling issue... rather than anything in the service config". Service feature flags are empty (no Express). Last successful deployment (2026-03-25) still runs and serves traffic.

Looks like corrupted deploy/scheduling state pinned to threset it?