1. **reset other msg in the a stream window**
The routine load job is a continuously scheduled job, and as the job
runs, previous errors do not need to be constantly displayed.

2. **show error info when  transaction of sub task failed**
If a subtask fails, it will continuously retry, and there may be some
errors that prevent the job from scheduling and consuming data properly,
such as continuous too many segments error(code: -235). At this time, it
is necessary to display it in a timely manner to make the user aware.

3. **set pause reason to other msg when reschedule job**
For jobs that are unexpectedly paused, the job manager has an auto
resume mechanism. However, for some scenarios, such as not being able to
connect to Kafka and being auto resumed after pause to retry, it may
cause users to not see the problem for a long time. Unexpectedly paused
jobs always have issues, even if auto resume occurs, the reason for the
error needs to be displayed.