Gateway: forward X-Fireworks-Genie header from client#63460
Conversation
Previously, there was no way to enable the "tracing" feature from Fireworks https://readme.fireworks.ai/docs/enabling-tracing This PR solves the problem by forwarding the `X-Fireworks-Genie` HTTP header to Fireworks if this HTTP header is set by the Gateway client. Fixes CODY-2555
eseliger
left a comment
There was a problem hiding this comment.
As talked about in person, this is free of charge, and a client side change will need to explicitly enable this for requests. Seems low risk 👍
|
|
||
| // Run the request transformer. | ||
| methods.transformRequest(req) | ||
| methods.transformRequest(r, req) |
There was a problem hiding this comment.
Might be nice to adopt the new naming scheme here as well (upstream and downstream req)
|
Post-hoc LGTM. This changes looks good. Although I wish we had better names than "upstream" and "downstream" for these concepts. I personally find it kinda confusing. (Thinking that the caller would be "upstream", and the thing I was forwarding the call to would be "downstream".)
|
|
@chrsmith agreed on naming. I went with upstream/downstream because I saw several mentions of "upstream" referring to the request going to the LLM provider. I didn't want to introduce a third naming convention. No objections to doing a batch rename to something clearer |
Previously, there was no way to enable the "tracing" feature from Fireworks https://readme.fireworks.ai/docs/enabling-tracing This PR solves the problem by forwarding the
X-Fireworks-GenieHTTP header to Fireworks if this HTTP header is set by the Gateway client.Fixes CODY-2555
Test plan
N/A
Changelog