ref(producers): clean up issues TaskProducers

[bmckerry]

Followup to #117081, #117107, #117360

TaskProducer is fully rolled out to the above producers in all regions. This PR removes the options controlling rollout and cleans up producer instantiation so we only have 1 producer running in a process (either SingletonProducer or TaskProducer based on what process is running).

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 33bdf82. Configure here.}

[cursor]

Import-time task producer selection

High Severity

_attribute_snapshot_producer is chosen once at import using settings.TASKWORKER_USE_TASK_PRODUCER. sentry.issues loads this module in AppConfig.ready() during configure(), before taskworker bootstrap sets that flag to True, so taskworker processes can keep a SingletonProducer for snapshot Kafka writes even when they should use TaskProducer.

 

^{Reviewed by Cursor Bugbot for commit 33bdf82. Configure here.}

[bmckerry]

This is valid, so I'll have to rethink how selecting the producer backend should work.
IMO ideally we'd just have a single producer we could use in both taskworker and non-taskworker processes.