Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix timeout and concurrency layer #51

Merged
merged 1 commit into from
Jan 3, 2024
Merged

Fix timeout and concurrency layer #51

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,6 +71,7 @@ tower-http = { version = "0.4.0", features = ["compression-br", "compression-gzi
rdkafka = { version = "0.33.2" }
axum-jrpc = { version = "0.5.1", features = ["serde_json", "anyhow_error"] }
ord-kafka-macros = { path = "ord-kafka-macros" }
tower = { version = "0.4.13", features = ["limit"] }


[dev-dependencies]
Expand Down
17 changes: 15 additions & 2 deletions src/subcommand/server.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,7 @@ use {
},
std::{cmp::Ordering, str, sync::Arc},
tokio_stream::StreamExt,
tower::limit::concurrency::ConcurrencyLimitLayer,
tower_http::{
compression::CompressionLayer,
cors::{Any, CorsLayer},
Expand Down Expand Up @@ -172,11 +173,17 @@ pub(crate) struct Server {
help = "Timeout requests after <SECONDS> seconds. Default: 30 seconds."
)]
timeout: Option<u64>,
#[clap(long, help = "Set max concurrent connections. Default: 1024")]
max_connections: Option<usize>,
}

impl Server {
pub(crate) fn run(self, options: Options, index: Arc<Index>, handle: Handle) -> SubcommandResult {
Runtime::new()?.block_on(async {
log::debug!(
"Starting server with {} max connections",
self.max_connections.unwrap_or(1024)
);
let index_clone = index.clone();

let index_thread = thread::spawn(move || loop {
Expand Down Expand Up @@ -276,7 +283,14 @@ impl Server {
.route("/tx/:txid", get(Self::transaction))

// API routes
.route("/rpc/v1", post(rpc::handler))
.route("/rpc/v1", post(rpc::handler)
.route_layer(TimeoutLayer::new(Duration::from_secs(self.timeout.unwrap_or(30))))
.route_layer(
ConcurrencyLimitLayer::new(
self.max_connections.unwrap_or(1024),
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Any reason for 1024? If so a comment would be nice

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

don't have a very concrete reason, just from the past we have only really seen 10 qps from the api side

)
)
)
.layer(axum::middleware::from_fn(middleware::tracing_layer))
.layer(Extension(index))
.layer(Extension(page_config))
Expand All @@ -295,7 +309,6 @@ impl Server {
.allow_origin(Any),
)
.layer(CompressionLayer::new())
.layer(TimeoutLayer::new(Duration::from_secs(self.timeout.unwrap_or(30))))
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

curious: why didn't this work?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not totally sure, I suspect that the very slow requests we saw were partially due to no concurrency limit, so there were just too many threads spawned and the timeout didn't trigger at the expected time

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I was able to see timeout (408 status) locally when picking a very low timeout, so it is definitely configured correctly

.with_state(server_config);

match (self.http_port(), self.https_port()) {
Expand Down
Loading