plugin-huawei: kill()/stop() not implemented on killable/pollable tasks

Sub-issue of kestra-io/kestra-ee#10964.

**Tasks:** io.kestra.plugin.huawei.dis.Consume, io.kestra.plugin.huawei.dms.kafka.Consume, io.kestra.plugin.huawei.dms.rocketmq.Consume

**Why:** Long-lived stream consume loops with no kill()/stop() override.

**Fix:** override `kill()` and/or `stop()` on these task classes to cancel the poll loop / close the consumer / cancel the remote job. See `io.kestra.plugin.gcp.bigquery.AbstractBigquery` for a reference implementation.

Locate io.kestra.plugin.huawei.dis.Consume, io.kestra.plugin.huawei.dms.kafka.Consume, and io.kestra.plugin.huawei.dms.rocketmq.Consume, then read io.kestra.plugin.gcp.bigquery.AbstractBigquery for the lifecycle reference. Verify how each long-lived consume loop, consumer, or remote job is stopped, and confirm that kill()/stop() cancels or closes the relevant work.

Written by the indexing model from the issue text.