The package-spec definition of routing rules doesn't explain what happens to these rules. As a developer I need to know how these affect the data stream behavior. IMO one of these sections should be expanded to include this information to developers.

My suggestion is to include this information:

Each rule generates an Elasticsearch reroute processor that is injected into the data stream's default ingest pipeline. The reroute processors are added after the injected "custom" pipeline processors (e.g. global@custom, <data_stream.type>@custom, <data_stream.type>-<package_name>.integration@custom, <data_stream.type>-<data_stream.dataset>@custom) such that the user's customizations run before the reroute rules.

References:

• https://github.com/elastic/kibana/blob/a493e4075b7267d34715607492197fc49c77b081/x-pack/plugins/fleet/server/services/epm/elasticsearch/ingest_pipeline/helpers.ts#L131-L136