huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-10 11:54:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

add metrics

Browse Source
This commit is contained in:
OlivierDehaene 2023-04-19 18:39:44 +02:00
parent ca98470cff
commit 521f6203d1
2 changed files with 7 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
router/src/infer.rs
View File

@ -446,7 +446,10 @@ fn filter_send_generations(generations: Vec<Generation>, entries: &mut IntMap<u6
// Send generation responses back to the infer task
// If the receive an error from the Flume channel, it means that the client dropped the
// request and we need to stop generating hence why we unwrap_or(true)
let stopped = send_responses(generation, entry).unwrap_or(true);
let stopped = send_responses(generation, entry).map_err(|err| {
metrics::increment_counter!("tgi_request_failure", "err" => "dropped");
err
}).unwrap_or(true);
if stopped {
entries.remove(&id).expect("ID not found in entries. This is a bug.");
}

4
router/src/queue.rs
View File

@ -161,6 +161,7 @@ impl State {
// Filter entries where the response receiver was dropped (== entries where the request
// was dropped by the client)
if entry.response_tx.is_disconnected() {
metrics::increment_counter!("tgi_request_failure", "err" => "dropped");
continue;
}
@ -190,6 +191,8 @@ impl State {
}
}
metrics::gauge!("tgi_queue_size", self.entries.len() as f64);
// Maybe all entries were dropped because their channel were closed
if batch_requests.is_empty() {
return None;
@ -207,7 +210,6 @@ impl State {
// Increment batch id
self.next_batch_id += 1;
metrics::gauge!("tgi_queue_size", self.entries.len() as f64);
metrics::histogram!("tgi_batch_next_size", batch.size as f64);
Some((batch_entries, batch, next_batch_span))
}