Wrap ContentTooLongException into a BatchSizeTooLargeException

[mpfz0r]

The dynamic batch size of archiving depends on catching these exceptions
and then retrying the scroll query with a smaller batch size.

Fixes https://github.com/Graylog2/graylog-plugin-enterprise/issues/3318

Tested with ES/OS containers with http.max_content_length=10mb env.

ES:

docker run --name elasticsearch -e "http.host=0.0.0.0" -p 9200:9200 -e "xpack.security.enabled=false" -e "http.max_content_length=10mb" -e "discovery.type=single-node" -e "action.auto_create_index=.watches,.triggered_watches,.watcher-history-*"   docker.elastic.co/elasticsearch/elasticsearch:7.10.2

OS:

docker run -d --name opensearch -p 9200:9200 -p 9600:9600 -e "http.max_content_length=10mb" -e "discovery.type=single-node" -e "DISABLE_INSTALL_DEMO_CONFIG=true" -e "DISABLE_SECURITY_PLUGIN=true" -e "action.auto_create_index=false" opensearchproject/opensearch:2.6.0

Ideally that should have an integration test, but that's just so much work :-/

[patrickmann]

Looks good - nice catch. Pretty elusive since it depends on document size.

On a side note: wondering why we don't have a base class for the search client classes. They have a lot of code in common.