Teleport uses the otelhttp library to automatically instrument our HTTP servers and clients with OTel tracing. Unfortunately, by default, it appears to use the entirety of the URL path when naming the spans. This makes it much harder to use the observability data and can cause performance issues with some observability platforms.
It would be preferable if instead each had a unique name that grouped requests based on their action, with the actual parameters being recorded as span attributes. For example, name the span DELETE /v2/thing/:thing_id rather than DELETE /v2/thing/aabbccdd.
This is backed up by the otel best practices
The span name concisely identifies the work represented by the Span, for example, an RPC method name, a function name, or the name of a subtask or stage within a larger computation. The span name SHOULD be the most general string that identifies a (statistically) interesting class of Spans, rather than individual Span instances while still being human-readable. That is, “get_user” is a reasonable name, while “get_user/314159”, where “314159” is a user ID, is not a good name due to its high cardinality. Generality SHOULD be prioritized over human-readability.
strideynet
commented
4 months ago
Teleport uses the
otelhttplibrary to automatically instrument our HTTP servers and clients with OTel tracing. Unfortunately, by default, it appears to use the entirety of the URL path when naming the spans. This makes it much harder to use the observability data and can cause performance issues with some observability platforms.It would be preferable if instead each had a unique name that grouped requests based on their action, with the actual parameters being recorded as span attributes. For example, name the span
DELETE /v2/thing/:thing_idrather thanDELETE /v2/thing/aabbccdd.This is backed up by the otel best practices