Skip to content

Migrate apollo_router_session_count_total metric to an OTel gauge #6495

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

Merged
goto-bus-stop merged 1 commit into from
Jan 7, 2025
Merged

Migrate apollo_router_session_count_total metric to an OTel gauge #6495

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
8 changes: 4 additions & 4 deletions apollo-router/src/axum_factory/axum_http_server_factory.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,16 +88,16 @@ fn session_count_instrument() -> ObservableGauge<u64> {
.init()
}

struct SessionCountGuard;
struct ActiveSessionCountGuard;

impl SessionCountGuard {
impl ActiveSessionCountGuard {
fn start() -> Self {
ACTIVE_SESSION_COUNT.fetch_add(1, Ordering::Acquire);
Self
}
}

impl Drop for SessionCountGuard {
impl Drop for ActiveSessionCountGuard {
fn drop(&mut self) {
ACTIVE_SESSION_COUNT.fetch_sub(1, Ordering::Acquire);
}
Expand Down Expand Up @@ -655,7 +655,7 @@ async fn handle_graphql(
experimental_log_on_broken_pipe: bool,
http_request: Request<DecompressionBody<Body>>,
) -> impl IntoResponse {
let _guard = SessionCountGuard::start();
let _guard = ActiveSessionCountGuard::start();

let (parts, body) = http_request.into_parts();

Expand Down
53 changes: 33 additions & 20 deletions apollo-router/src/axum_factory/listeners.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,8 @@ use futures::channel::oneshot;
use futures::prelude::*;
use hyper::server::conn::Http;
use multimap::MultiMap;
use opentelemetry::metrics::MeterProvider;
use opentelemetry::KeyValue;
#[cfg(unix)]
use tokio::net::UnixListener;
use tokio::sync::mpsc;
Expand All @@ -26,13 +28,29 @@ use crate::axum_factory::ENDPOINT_CALLBACK;
use crate::configuration::Configuration;
use crate::http_server_factory::Listener;
use crate::http_server_factory::NetworkStream;
use crate::metrics::meter_provider;
use crate::router::ApolloRouterError;
use crate::router_factory::Endpoint;
use crate::ListenAddr;

static SESSION_COUNT: AtomicU64 = AtomicU64::new(0);
static TOTAL_SESSION_COUNT: AtomicU64 = AtomicU64::new(0);
static MAX_FILE_HANDLES_WARN: AtomicBool = AtomicBool::new(false);

struct TotalSessionCountGuard;

impl TotalSessionCountGuard {
fn start() -> Self {
TOTAL_SESSION_COUNT.fetch_add(1, Ordering::Acquire);
Self
}
}

impl Drop for TotalSessionCountGuard {
fn drop(&mut self) {
TOTAL_SESSION_COUNT.fetch_sub(1, Ordering::Acquire);
}
}

#[derive(Clone, Debug)]
pub(crate) struct ListenAddrAndRouter(pub(crate) ListenAddr, pub(crate) Router);

Expand Down Expand Up @@ -214,9 +232,19 @@ pub(super) fn serve_router_on_listen_addr(
let server = async move {
tokio::pin!(shutdown_receiver);

let connection_shutdown = Arc::new(Notify::new());
let _total_session_count_instrument = meter_provider()
.meter("apollo/router")
.u64_observable_gauge("apollo_router_session_count_total")
.with_description("Number of currently connected clients")
.with_callback(move |gauge| {
gauge.observe(
TOTAL_SESSION_COUNT.load(Ordering::Relaxed),
&[KeyValue::new("listener", address.to_string())],
);
})
.init();

let address = address.to_string();
let connection_shutdown = Arc::new(Notify::new());

loop {
tokio::select! {
Expand All @@ -234,16 +262,10 @@ pub(super) fn serve_router_on_listen_addr(
tracing::info!("can accept connections again");
MAX_FILE_HANDLES_WARN.store(false, Ordering::SeqCst);
}

// We only want to recognise sessions if we are the main graphql port.
if main_graphql_port {
let session_count = SESSION_COUNT.fetch_add(1, Ordering::Acquire)+1;
tracing::info!(
value.apollo_router_session_count_total = session_count,
listener = &address
);
}
let _guard = main_graphql_port.then(TotalSessionCountGuard::start);
Copy link
Member Author

@goto-bus-stop goto-bus-stop Jan 13, 2025

Choose a reason for hiding this comment

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

Note this contains a mistake. The guard must be moved into the task::spawn call below, but it isn't, so the total session count metric is broken with this PR (it is immediately decremented before the next incoming request is handled.)

#6527 will fix this problem.

Copy link
Member Author

@goto-bus-stop goto-bus-stop Jan 13, 2025

Choose a reason for hiding this comment

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

That PR will be redone in a different way. #6539 fixes just this problem.


let address = address.clone();
let mut http_config = http_config.clone();
tokio::task::spawn(async move {
// this sender must be moved into the session to track that it is still running
Expand Down Expand Up @@ -352,15 +374,6 @@ pub(super) fn serve_router_on_listen_addr(
}
}
}

// We only want to recognise sessions if we are the main graphql port.
if main_graphql_port {
let session_count = SESSION_COUNT.fetch_sub(1, Ordering::Acquire)-1;
tracing::info!(
value.apollo_router_session_count_total = session_count,
listener = &address
);
}
});
}

Expand Down