Skip to content

Give SDK work a predictable runtime and shutdown #127

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

Open
Stephen Belanger (Qard) wants to merge 2 commits into
base: refactor/finished-span-lifecycle
Choose a base branch
from
+1,803 −296
Open

Give SDK work a predictable runtime and shutdown #127

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6d18a20
Bind SDK work to an explicit or owned runtime
Qard Oct 20, 2026
3b254db
Coordinate response tasks and delivery during shutdown
Qard Oct 20, 2026
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
4 changes: 4 additions & 0 deletions src/error.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,10 @@ use thiserror::Error;

#[derive(Debug, Error)]
pub enum BraintrustError {
#[error(transparent)]
Execution(#[from] crate::ExecutionError),
#[error(transparent)]
Admission(#[from] crate::EnqueueError),
#[error("invalid configuration: {0}")]
InvalidConfig(String),
#[error("http error: {0}")]
Expand Down
137 changes: 84 additions & 53 deletions src/internal/delivery.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,13 +13,13 @@ use std::{
time::Duration,
};
use tokio::{
runtime::Handle,
sync::{oneshot, Notify, Semaphore},
task::AbortHandle,
};
use tracing::warn;

use crate::internal::write_outcome::{WriteOutcome, WriteResponder};
use crate::internal::execution::ExecutionBinding;
use crate::internal::shared_outcome::{WriteOutcome, WriteResponder};
type WriteSender = WriteResponder;
type FlushSender = oneshot::Sender<Result<FlushReport, FlushError>>;
#[derive(Clone)]
Expand Down Expand Up @@ -215,7 +215,7 @@ impl<C> Inner<C> {
pub(crate) struct WriteService<C> {
inner: Arc<Inner<C>>,
owners: Arc<ProducerOwner<C>>,
runtime: Handle,
runtime: ExecutionBinding,
exit: Shared<BoxFuture<'static, Result<(), ExecutionError>>>,
}
impl<C> Clone for WriteService<C> {
Expand All @@ -239,7 +239,7 @@ impl<C> Drop for ProducerOwner<C> {
}
}
impl<C: Send + 'static> WriteService<C> {
pub fn new<F>(capacity: usize, runtime: Handle, process: F) -> Self
pub fn new<F>(capacity: usize, runtime: ExecutionBinding, process: F) -> Self
where
F: Fn(Vec<C>) -> BoxFuture<'static, anyhow::Result<()>> + Send + 'static,
{
Expand Down Expand Up @@ -394,6 +394,13 @@ impl<C: Send + 'static> WriteService<C> {
completion
}

pub fn cancel(&self, error: ExecutionError) {
self.inner.stop(error);
if let Some(abort) = lock(&self.inner.abort).as_ref() {
abort.abort();
}
}

#[cfg(test)]
pub fn is_empty(&self) -> bool {
lock(&self.inner.state).outstanding == 0
Expand Down Expand Up @@ -472,15 +479,19 @@ mod tests {
let (entered, receiver) = mpsc::unbounded_channel();
let gate = Arc::new(Semaphore::new(0));
let process_gate = gate.clone();
let service = WriteService::new(capacity, Handle::current(), move |commands| {
let gate = process_gate.clone();
let entered = entered.clone();
Box::pin(async move {
entered.send(commands).unwrap();
gate.acquire().await.unwrap().forget();
Ok(())
})
});
let service = WriteService::new(
capacity,
ExecutionBinding::external(tokio::runtime::Handle::current()),
move |commands| {
let gate = process_gate.clone();
let entered = entered.clone();
Box::pin(async move {
entered.send(commands).unwrap();
gate.acquire().await.unwrap().forget();
Ok(())
})
},
);
(service, receiver, gate)
}

Expand Down Expand Up @@ -516,18 +527,22 @@ mod tests {
let (entered, mut receiver) = mpsc::unbounded_channel();
let gate = Arc::new(Semaphore::new(0));
let process_gate = gate.clone();
let service = WriteService::new(4, Handle::current(), move |commands: Vec<usize>| {
let entered = entered.clone();
let gate = process_gate.clone();
Box::pin(async move {
entered.send(commands.clone()).unwrap();
gate.acquire().await.unwrap().forget();
if commands == [1] {
anyhow::bail!("first batch failed");
}
Ok(())
})
});
let service = WriteService::new(
4,
ExecutionBinding::external(tokio::runtime::Handle::current()),
move |commands: Vec<usize>| {
let entered = entered.clone();
let gate = process_gate.clone();
Box::pin(async move {
entered.send(commands.clone()).unwrap();
gate.acquire().await.unwrap().forget();
if commands == [1] {
anyhow::bail!("first batch failed");
}
Ok(())
})
},
);
let first = service.enqueue(1);
assert_eq!(receiver.recv().await.unwrap(), vec![1]);
let second = service.enqueue(2);
Expand All @@ -548,11 +563,15 @@ mod tests {

#[tokio::test]
async fn graceful_shutdown_reports_its_failed_batch_after_notifying_all_writes() {
let service = WriteService::new(4, Handle::current(), |_: Vec<usize>| {
Box::pin(async {
anyhow::bail!("terminal failure");
})
});
let service = WriteService::new(
4,
ExecutionBinding::external(tokio::runtime::Handle::current()),
|_: Vec<usize>| {
Box::pin(async {
anyhow::bail!("terminal failure");
})
},
);
let first = service.enqueue(1);
let second = service.enqueue(2);
let shutdown = service.shutdown(Duration::from_secs(1));
Expand Down Expand Up @@ -605,15 +624,19 @@ mod tests {
async fn every_failed_batch_notifies_all_members_releases_errors_and_allows_fresh_work() {
let fail = Arc::new(AtomicBool::new(true));
let process_fail = fail.clone();
let service = WriteService::new(4, Handle::current(), move |_: Vec<usize>| {
let fail = process_fail.load(Ordering::Relaxed);
Box::pin(async move {
if fail {
anyhow::bail!("terminal failure");
}
Ok(())
})
});
let service = WriteService::new(
4,
ExecutionBinding::external(tokio::runtime::Handle::current()),
move |_: Vec<usize>| {
let fail = process_fail.load(Ordering::Relaxed);
Box::pin(async move {
if fail {
anyhow::bail!("terminal failure");
}
Ok(())
})
},
);
for _ in 0..100 {
let first = service.enqueue(1);
let second = service.enqueue(2);
Expand Down Expand Up @@ -765,15 +788,19 @@ mod tests {
let (entered, mut receiver) = mpsc::unbounded_channel();
let gate = Arc::new(Semaphore::new(0));
let process_gate = gate.clone();
let service = WriteService::new(4, Handle::current(), move |_: Vec<usize>| {
let entered = entered.clone();
let gate = process_gate.clone();
Box::pin(async move {
entered.send(()).unwrap();
gate.acquire().await.unwrap().forget();
panic!("test processor failed");
})
});
let service = WriteService::new(
4,
ExecutionBinding::external(tokio::runtime::Handle::current()),
move |_: Vec<usize>| {
let entered = entered.clone();
let gate = process_gate.clone();
Box::pin(async move {
entered.send(()).unwrap();
gate.acquire().await.unwrap().forget();
panic!("test processor failed");
})
},
);
let active = service.enqueue(1);
receiver.recv().await.unwrap();
let pending = service.enqueue(2);
Expand Down Expand Up @@ -817,11 +844,15 @@ mod tests {
let processed = Arc::new(std::sync::atomic::AtomicUsize::new(0));
let process_seen = seen.clone();
let process_count = processed.clone();
let service = WriteService::new(1024, Handle::current(), move |commands: Vec<usize>| {
lock(&process_seen).extend(commands.iter().copied());
process_count.fetch_add(commands.len(), Ordering::Relaxed);
Box::pin(async { Ok(()) })
});
let service = WriteService::new(
1024,
ExecutionBinding::external(tokio::runtime::Handle::current()),
move |commands: Vec<usize>| {
lock(&process_seen).extend(commands.iter().copied());
process_count.fetch_add(commands.len(), Ordering::Relaxed);
Box::pin(async { Ok(()) })
},
);
let (writes, flushes) = std::thread::scope(|scope| {
let producers: Vec<_> = (0..4)
.map(|producer| {
Expand Down
Loading
Loading