Version the worker-bound TaskInstance fields in the execution API schema

[kaxil]

Why

Since #65958 the supervisor routes every task through get_coordinator_manager().for_queue(ti.queue), so it depends on queue being on the task instance it receives. That field (along with pool_slots and priority_weight) lived in a BaseTaskInstanceDTO duplicated between airflow-core and task-sdk (#67174), kept aligned by an AST-comparison prek hook, while the TaskInstance schema in the versioned execution API spec lacked it.

That spec is already the designated home for this vocabulary: get_extra_schemas() injects TaskInstance (and BundleInfo, imported from airflow.executors.workloads) into the OpenAPI schema precisely so client SDKs generate these types from a versioned source. Adding the missing field there gives one source of truth instead of two hand-synced copies.

What

• Add queue to the execution API TaskInstance schema, with an AddTaskInstanceQueueField Cadwyn version change in the 2026-06-30 bundle, and regenerate the task-sdk datamodels.
• Re-point StartupDetails.ti, supervise_task() and the coordinator interfaces at the generated TaskInstance.
• Fold the core TaskInstanceDTO into a subclass of the schema model that adds only the executor-side fields: pool_slots and priority_weight (read by the executor's priority sort and the edge executor's concurrency accounting, never by the worker), plus external_executor_id and executor_config (excluded from serialization).
• Delete the duplicated task-sdk execution_time.workloads module and the check-task-instance-dto-sync hook. A round-trip test (executor DTO JSON validated by the SDK's generated model) covers the contract the hook used to enforce.
• Remove parent_context_carrier from the DTO: it has no Python readers or writers anywhere in core, task-sdk or providers.

Design notes

• Only queue is worker-facing: the supervisor routes on it. pool_slots and priority_weight are executor concerns and stay on the executor DTO, still serialized in the workload (required, exactly the pre-PR wire shape) so older workers that deserialize the workload with the required-field model keep working.
• queue gets a default ("default") instead of being required. The executor always sends the real value (ExecuteTask.make validates from the ORM task instance, where the column is non-null); the default keeps hand-built instances such as tests and dry_run runs valid, and matches how map_index and hostname are already modelled in this schema.
• No released artifact ships the deleted surfaces: the duplicated DTO module and the supervisor wire-schema bundle only exist on the unreleased 3.3 line, so no compat shims are needed. The supervisor wire-schema snapshot change rides the in-progress floor version of that bundle, where no VersionChange entry is possible.
• The edge3 worker API spec regenerates via its existing hook because it embeds the workload TI shape. The checked-in TypeScript/Go clients under plugins/www/openapi-gen and go-sdk/pkg/edgeapi are not regenerated here: their codegen is not wired into prek and already lags the spec on main (the Go client was last regenerated in Simple sending of return value from Go tasks to XCom #56481), and the affected fields are optional in both.

[ashb]

through get_coordinator_manager().for_queue(ti.queue), so it depends on queue, pool_slots and priority_weight on the task instance it receives.

The last two being required on the worker seems like a code smell. If it makes it to the worker it should just executed. That is the scheduler's job, not the workers

[kaxil]

Agreed and fixed in 78a6b82 -- the worker-facing schema now only gains queue; pool_slots and priority_weight remain executor-only fields on the DTO.

[ashb]

Few nits/Qs but LGTM overall

[jason810496]

Really nice catch, thanks. Do we need full tests before merge Btw?