Celery group nested within chain does not call chain's link_error task on failure

I have the following workflow triggered from within a separate task on Celery 5.1.2:

    work_flow = chain([
        task_1.si(args),
        data_group,
        task_2.si(args),
        task_3.si(args),
        task_4.si(args),
    ]).apply_async(link_error=clean_up.s(args))

The variable data_group is defined beforehand:

    data_group = group([
        task_a.si(args),
        task_b.si(args),
    ])

My issue is that failing task_a or task_b using AsyncResult's revoke() method and the SIGUSR1 signal (which raises a SoftTimeLimitExceeded exception during execution) does not seem to trigger the clean_up task.

I have gone through my Celery logs and there are two warnings aside from the SoftTimeLimitExceeded exception which I induced:

[WARNING/MainProcess] Can't find ChordCounter for Group

[WARNING/ForkPoolWorker-3] Can't find ChordCounter for Group

I have been looking for similar issues online and reading through the Canvas: Designing Workflows section of the Celery documentation. I saw in the docs that "the return values of a group are not collected to be passed to a linked callback signature" (https://docs.celeryq.dev/en/stable/userguide/canvas.html#group-callbacks-and-error-handling) and was wondering if this would explain my issue. In that case, I am not sure how to move forward and restructure my workflow such that tasks a and b can both be run in parallel and call my clean_up task on failure.