Define the routing connector in the Collector config with a table mapping attribute values to named exporters or pipelines
Set a default pipeline to handle telemetry that does not match any routing rule
Use resource or span attributes such as service.name or deployment.environment as the routing key
Connect the routing connector as both a receiver in downstream pipelines and an exporter in the upstream pipeline
Validate that telemetry reaches the correct downstream destination by checking exporter metrics for each route
Known gotchas
The routing connector is evaluated per data item, not per batch; high fan-out routing tables can add CPU overhead to high-throughput pipelines
If the routing attribute is missing from some telemetry, ensure the default route is configured or those items will be dropped silently
Routing connector is a connector component, not a processor; it must be placed at the boundary between two pipeline definitions rather than inline in one pipeline
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp