optimize: use elasticsearch msearch API for batch queries

[KANIOYH]

This commit refactors the Elasticsearch batch_query
to utilize the msearch (multi-search) API instead of
sending individual queries in a loop.

The msearch API allows sending multiple search requests in a single
HTTP call, which significantly reduces network overhead and improves
query throughput, especially when performing batch operations.

Changes include:

• Increased Elasticsearch thread_pool.search.queue_size in Dockerfile
• Replaced batch_query calls with msearch API in module.py

Attention：

• Set thread_pool.search.queue_size=1000 (with thread_pool.search.size=1),
  allowing more requests to queue but still executing them in a single thread