Scalable push query should terminate if the underlying persistent query is terminated

[nateab]

Describe the bug
When you terminate a persistent query, any scalable push query that was based on it should also be terminated since there will be no additional updates to receive, since scalable push queries don't read from the underlying kafka topic unlike regular push queries.

To Reproduce
Steps to reproduce the behavior, include:

1. The version of KSQL.
   0.20

2. Sample source data.

3. Any SQL statements you ran

CREATE STREAM RATINGS (RATING_ID BIGINT, USER_ID INTEGER, STARS INTEGER, ROUTE_ID BIGINT KEY, RATING_TIME BIGINT, CHANNEL STRING, MESSAGE STRING) WITH (KAFKA_TOPIC='ratings', VALUE_FORMAT='JSON', partitions=1);

CREATE STREAM RATINGS_STREAM AS SELECT CAST((RANDOM() * 1000) AS INTEGER) KEY, * FROM RATINGS RATINGS;

SELECT * FROM RATINGS_STREAM EMIT CHANGES;
TERMINATE CSAS_RATINGS_STREAM_3;

Expected behavior
The persistent query is terminated, and the push query running on the persistent query is also terminated.
Actual behaviour
A clear and concise description of what actually happens, including:

1. CLI output
2. Error messages
3. KSQL logs

The persistent query does terminate, but the scalable push query continues to run even though it will not report any changes.

Additional context
Add any other context about the problem here.

[PeterLindner]

I disagree with the assumption that no additional updates will be recieved when the persistent query is terminated.
At any point in time another persistent query, which inserts into the stream, could be started or a application outside of ksqlDB could produce into the underlying kafka topic.
I firmly believe that consuming from a stream / topic should be decoupled from producing into it.

[nateab]

Yes these are fair points, I misunderstood when I created the issue, but this is specifically for scalable push queries. It would be good in the future to also decouple consuming and producing but the current implementation of scalable push queries does not yet allow for this.

[agavra]

might be related to #3967