[FEATURE] Add limitations settings for datasource, query, sessions in domain level instead of datasource

[penghuo]

• Max number of datasources per domain — 20

  • plugins.query.executionengine.spark.datasources.limit

• Max number of refresh job of (skipping/covering/materialized views) per domain — 50

  • plugins.query.executionengine.spark.refresh_job.limit

• Max number of concurrent sessions (for REPL jobs) per datasource - 100 (Existing feature)

  • plugins.query.executionengine.spark.session.limit

• Max number of batch queries (in case if REPL is disabled)

  • we disable batch query, no need to add batch query limitation.

[penghuo]

Stage - I

Implement LeaseManagerService in Plugin.

• QueryHandler borrow lease before execute the query.
• LeaseManagerService check request against domain level concurrent limit.
• LeaseManagerService running on data node and check limit based on cluster settings.

API

• void LeaseManagerService.borrow(BorrowLeaseRequest request)
  • ConcurrencyLimitExceededException

Model

• BorrowLeaseRequest
  • QueryType
  • DataSourceName