[FEATURE] Add maxExecutors configuration for streaming queries

[penghuo]

Is your feature request related to a problem?
Currently, both interactive and streaming queries share configurations for Spark executors. When streaming jobs are running, executors are dynamically allocated with an upper limit set by spark.dynamicAllocation.maxExecutors (the default value is 30). Spark's dynamic resource allocation (DRA) system aggressively allocates more executors than necessary to achieve low latency, leading to inefficiencies, because the volume of data processed in each interval is relatively small (100GB per 15 minutes), and the latency requirement is not stringent (15 minutes).

What solution would you like?
add spark.dynamicAllocation.streaming.maxExecutors configuration and default value is 10

What alternatives have you considered?
in openserach-sql, submit interactive and streaming query separately and configure spark.dynamicAllocation.maxExecutors.