apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-07-23 16:51:04 +02:00
Code Issues Projects Releases Packages Wiki Activity

fix start time of llm filter

Browse source
This commit is contained in:
Adil Hafeez 2024-11-17 16:43:21 -08:00
parent d3c17c7abd
commit 3e6786a1df
1 changed files with 8 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

9
crates/llm_gateway/src/stream_context.rs
View file

@ -40,6 +40,7 @@ pub struct StreamContext {
ttft_duration: Option<Duration>, ttft_duration: Option<Duration>,
ttft_time: Option<SystemTime>, ttft_time: Option<SystemTime>,
pub traceparent: Option<String>, pub traceparent: Option<String>,
request_body_sent_time: Option<SystemTime>,
user_message: Option<Message>, user_message: Option<Message>,
} }
@ -60,6 +61,7 @@ impl StreamContext {
traceparent: None, traceparent: None,
ttft_time: None, ttft_time: None,
user_message: None, user_message: None,
request_body_sent_time: None,
} }
} }
fn llm_provider(&self) -> &LlmProvider { fn llm_provider(&self) -> &LlmProvider {
@ -196,6 +198,11 @@ impl HttpContext for StreamContext {
fn on_http_request_body(&mut self, body_size: usize, end_of_stream: bool) -> Action { fn on_http_request_body(&mut self, body_size: usize, end_of_stream: bool) -> Action {
// Let the client send the gateway all the data before sending to the LLM_provider. // Let the client send the gateway all the data before sending to the LLM_provider.
// TODO: consider a streaming API. // TODO: consider a streaming API.
if self.request_body_sent_time.is_none() {
self.request_body_sent_time = Some(get_current_time().unwrap());
}
if !end_of_stream { if !end_of_stream {
return Action::Pause; return Action::Pause;
} }
@ -351,7 +358,7 @@ impl HttpContext for StreamContext {
"upstream_llm_time".to_string(), "upstream_llm_time".to_string(),
parent_trace_id.to_string(), parent_trace_id.to_string(),
Some(parent_span_id.to_string()), Some(parent_span_id.to_string()),
self.start_time self.request_body_sent_time
.unwrap() .unwrap()
.duration_since(UNIX_EPOCH) .duration_since(UNIX_EPOCH)
.unwrap() .unwrap()