Managed deploy upgrades stopped crashing on the pre-upgrade migration hook

Fixes a crash that wedged managed Tuist deployments during helm upgrade. The pre-upgrade migration hook was failing to start because the migrate Job’s S3 boot environment variables (TUIST_S3_BUCKET_NAME, TUIST_S3_ENDPOINT, TUIST_S3_REGION, TUIST_S3_VIRTUAL_HOST, TUIST_S3_BUCKET_AS_HOST) had been gated behind not managedSecrets, while the server Deployment already set them. With the encrypted secrets blob removed, the migrate Job received no S3 endpoint and runtime.exs crashed with URI.parse(nil), hanging the whole deploy cascade. The chart now sets the S3 boot config on the migrate Job unconditionally, so managed deployments boot the migration hook and the upgrade cascade completes. Self-hosted installs see no change in behavior.