mirrors/activitypub-federation-rust
1
0
Fork 0
mirror of https://github.com/LemmyNet/activitypub-federation-rust.git synced 2024-06-08 08:19:28 +00:00
Code Issues Projects Releases Wiki Activity

Expose the worker & tweak the ActivityQueue trait

Browse source
This commit is contained in:
cetra3 2023-07-11 18:46:12 +09:30
parent 57d5b86370
commit 5bd0bf0a2a
6 changed files with 217 additions and 179 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

60
src/config.rs
View file

@ -18,7 +18,12 @@
use crate::{
error::Error,
protocol::verification::verify_domains_match,
queue::{simple_queue::SimpleQueue, ActivityQueue, SendActivityTask},
queue::{
standard_retry_queue::StandardRetryQueue,
worker::Worker,
ActivityMessage,
ActivityQueue,
},
traits::{ActivityHandler, Actor},
};
use async_trait::async_trait;
@ -96,7 +101,7 @@ pub struct FederationConfig<T: Clone> {
/// Queue for sending outgoing activities. Only optional to make builder work, its always
/// present once constructed.
#[builder(setter(custom))]
pub(crate) activity_queue: Sender<SendActivityTask>,
pub(crate) activity_queue: Option<Sender<ActivityMessage>>,
}
impl<T: Clone> FederationConfig<T> {
@ -205,17 +210,7 @@ impl<T: Clone> FederationConfigBuilder<T> {
&mut self,
queue: Q,
) -> &mut Self {
let (sender, mut receiver) = channel(8192);
tokio::spawn(async move {
while let Some(message) = receiver.recv().await {
if let Err(err) = queue.queue(message).await {
error!("{err:?}");
}
}
});
self.activity_queue = Some(sender);
self.activity_queue = Some(Some(make_sender(queue)));
self
}
@ -225,25 +220,38 @@ impl<T: Clone> FederationConfigBuilder<T> {
/// queue for outgoing activities, which is stored internally in the config struct.
/// Requires a tokio runtime for the background queue.
pub async fn build(&mut self) -> Result<FederationConfig<T>, FederationConfigBuilderError> {
if self.activity_queue.is_none() {
let queue = SimpleQueue::new(
self.client
.clone()
.unwrap_or_else(|| reqwest::Client::default().into()),
self.worker_count.unwrap_or_default(),
self.retry_count.unwrap_or_default(),
self.request_timeout.unwrap_or(Duration::from_secs(10)),
60,
self.http_signature_compat.unwrap_or_default(),
);
let mut config = self.partial_build()?;
self.activity_queue(queue).await;
if config.activity_queue.is_none() {
if config.debug {
// If we're in debug we use a single worker to send things
let queue = Worker::from_config(&config);
config.activity_queue = Some(make_sender(queue));
} else {
// Otherwise we use the standard retry queue
let queue = StandardRetryQueue::from_config(&config);
config.activity_queue = Some(make_sender(queue));
}
}
self.partial_build()
Ok(config)
}
}
fn make_sender<Q: ActivityQueue + Sync + Send + 'static>(queue: Q) -> Sender<ActivityMessage> {
let (sender, mut receiver) = channel(8192);
tokio::spawn(async move {
while let Some(message) = receiver.recv().await {
if let Err(err) = queue.queue(message).await {
error!("{err:?}");
}
}
});
sender
}
// impl<T: Clone> FederationConfig<T> {
// /// Shut down this federation, waiting for the outgoing queue to be sent.
// /// If the activityqueue is still in use in other requests or was never constructed, returns an error.

32
src/queue/mod.rs
View file

@ -3,9 +3,9 @@
#![doc = include_str!("../../docs/09_sending_activities.md")]
pub(crate) mod request;
pub mod simple_queue;
mod util;
mod worker;
pub mod standard_retry_queue;
pub mod util;
pub mod worker;
use crate::{
config::Data,
@ -36,11 +36,8 @@ pub trait ActivityQueue {
/// The errors that can be returned when queuing
type Error: Debug;
/// Retrieve the queue stats
fn stats(&self) -> &Stats;
/// Queues one activity task to a specific inbox
async fn queue(&self, message: SendActivityTask) -> Result<(), Self::Error>;
async fn queue(&self, message: ActivityMessage) -> Result<(), Self::Error>;
}
/// Sends an activity with an outbound activity queue
@ -57,7 +54,11 @@ where
ActorType: Actor,
{
let config = &data.config;
let queue = data.config.activity_queue.clone();
let queue = data
.config
.activity_queue
.clone()
.expect("Activity Queue is always configured");
let actor_id = activity.actor();
let activity_id = activity.id();
let activity_serialized: Bytes = serde_json::to_vec(&activity)?.into();
@ -85,7 +86,7 @@ where
continue;
}
let message = SendActivityTask {
let message = ActivityMessage {
id: Uuid::new_v4(),
actor_id: actor_id.clone(),
activity_id: activity_id.clone(),
@ -129,7 +130,7 @@ where
#[derive(Clone, Debug)]
/// The struct sent to a worker for processing
/// When `send_activity` is used, it is split up to tasks per-inbox
pub struct SendActivityTask {
pub struct ActivityMessage {
/// The ID of the activity task
pub id: Uuid,
/// The actor ID
@ -160,7 +161,7 @@ impl Debug for Stats {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(
f,
"Activity queue stats: pending: {}, running: {}, retries: {}, dead: {}, complete: {}",
"Activity queue stats: pending: {}, running: {}, retries: {}, dead (last hr): {}, complete (last hr): {}",
self.pending.load(Ordering::Relaxed),
self.running.load(Ordering::Relaxed),
self.retries.load(Ordering::Relaxed),
@ -180,7 +181,10 @@ mod tests {
time::{Duration, Instant},
};
use crate::{http_signatures::generate_actor_keypair, queue::simple_queue::SimpleQueue};
use crate::{
http_signatures::generate_actor_keypair,
queue::standard_retry_queue::StandardRetryQueue,
};
use tracing::{debug, info};
use super::*;
@ -237,7 +241,7 @@ mod tests {
*/
let queue = SimpleQueue::new(
let queue = StandardRetryQueue::new(
reqwest::Client::default().into(),
num_workers,
num_workers,
@ -248,7 +252,7 @@ mod tests {
let keypair = generate_actor_keypair().unwrap();
let message = SendActivityTask {
let message = ActivityMessage {
id: Uuid::new_v4(),
actor_id: "http://localhost:8001".parse().unwrap(),
activity_id: "http://localhost:8001/activity".parse().unwrap(),

28
src/queue/request.rs
View file

@ -16,10 +16,12 @@ use crate::{
use anyhow::{anyhow, Context};
use tracing::debug;
use super::{util::RetryStrategy, SendActivityTask};
use super::{util::RetryStrategy, ActivityMessage};
pub(super) async fn sign_and_send(
task: &SendActivityTask,
/// Sign and send a message with an optional retry
/// The retry itself doesn't re-sign the request, so the retry times should be < 5 min
pub async fn sign_and_send(
message: &ActivityMessage,
client: &ClientWithMiddleware,
timeout: Duration,
retry_strategy: RetryStrategy,
@ -27,19 +29,19 @@ pub(super) async fn sign_and_send(
) -> Result<(), anyhow::Error> {
debug!(
"Sending {} to {}, contents:\n {}",
task.activity_id,
task.inbox,
serde_json::from_slice::<serde_json::Value>(&task.activity)?
message.activity_id,
message.inbox,
serde_json::from_slice::<serde_json::Value>(&message.activity)?
);
let request_builder = client
.post(task.inbox.to_string())
.post(message.inbox.to_string())
.timeout(timeout)
.headers(generate_request_headers(&task.inbox));
.headers(generate_request_headers(&message.inbox));
let request = sign_request(
request_builder,
&task.actor_id,
task.activity.clone(),
task.private_key.clone(),
&message.actor_id,
message.activity.clone(),
message.private_key.clone(),
http_signature_compat,
)
.await
@ -48,7 +50,7 @@ pub(super) async fn sign_and_send(
retry(
|| {
send(
task,
message,
client,
request
.try_clone()
@ -61,7 +63,7 @@ pub(super) async fn sign_and_send(
}
pub(super) async fn send(
task: &SendActivityTask,
task: &ActivityMessage,
client: &ClientWithMiddleware,
request: Request,
) -> Result<(), anyhow::Error> {

136
src/queue/simple_queue.rs → src/queue/standard_retry_queue.rs
View file

@ -11,45 +11,56 @@ use tokio::{
sync::mpsc::{unbounded_channel, UnboundedSender},
task::{JoinHandle, JoinSet},
};
use tracing::{info, warn};
use crate::config::FederationConfig;
use super::{
util::RetryStrategy,
worker::{activity_worker, retry_worker},
util::{retry, RetryStrategy},
worker::Worker,
ActivityMessage,
ActivityQueue,
SendActivityTask,
Stats,
};
/// A simple activity queue which spawns tokio workers to send out requests
/// When creating a queue, it will spawn a task per worker thread
/// Uses an unbounded mpsc queue for communication (i.e, all messages are in memory)
pub struct SimpleQueue {
/// A tokio spawned worker queue which is responsible for submitting requests to federated servers
/// This will retry up to one time with the same signature, and if it fails, will move it to the retry queue.
/// We need to retry activity sending in case the target instances is temporarily unreachable.
/// In this case, the task is stored and resent when the instance is hopefully back up. This
/// list shows the retry intervals, and which events of the target instance can be covered:
/// - 60s (one minute, service restart) -- happens in the worker w/ same signature
/// - 60min (one hour, instance maintenance) --- happens in the retry queue
/// - 60h (2.5 days, major incident with rebuild from backup) --- happens in the retry queue
pub struct StandardRetryQueue {
// Stats shared between the queue and workers
stats: Arc<Stats>,
sender: UnboundedSender<SendActivityTask>,
worker_count: usize,
sender: UnboundedSender<ActivityMessage>,
sender_task: JoinHandle<()>,
retry_sender_task: JoinHandle<()>,
}
#[async_trait]
impl ActivityQueue for SimpleQueue {
impl ActivityQueue for StandardRetryQueue {
type Error = anyhow::Error;
fn stats(&self) -> &Stats {
&self.stats
}
async fn queue(&self, message: SendActivityTask) -> Result<(), Self::Error> {
async fn queue(&self, message: ActivityMessage) -> Result<(), Self::Error> {
self.stats.pending.fetch_add(1, Ordering::Relaxed);
self.sender.send(message)?;
let running = self.stats.running.load(Ordering::Relaxed);
if running == self.worker_count && self.worker_count != 0 {
warn!("Reached max number of send activity workers ({}). Consider increasing worker count to avoid federation delays", self.worker_count);
warn!("{:?}", self.stats);
} else {
info!("{:?}", self.stats);
}
Ok(())
}
}
impl SimpleQueue {
/// Construct a queue from federation config
impl StandardRetryQueue {
/// Builds a simple queue from a federation config
pub fn from_config<T: Clone>(config: &FederationConfig<T>) -> Self {
Self::new(
config.client.clone(),
@ -85,7 +96,6 @@ impl SimpleQueue {
let (retry_sender, mut retry_receiver) = unbounded_channel();
let retry_stats = stats.clone();
let retry_client = client.clone();
// The "fast path" retry
// The backoff should be < 5 mins for this to work otherwise signatures may expire
@ -97,6 +107,13 @@ impl SimpleQueue {
initial_sleep: 0,
};
let worker = Arc::new(Worker {
client: client.clone(),
request_timeout: timeout,
strategy,
http_signature_compat,
});
// The "retry path" strategy
// After the fast path fails, a task will sleep up to backoff ^ 2 and then retry again
let retry_strategy = RetryStrategy {
@ -106,17 +123,23 @@ impl SimpleQueue {
initial_sleep: backoff.pow(2), // wait 60 mins before even trying
};
let retry_worker = Arc::new(Worker {
client: client.clone(),
request_timeout: timeout,
// Internally we need to re-sign the message each attempt so we remove this strategy
strategy: RetryStrategy::default(),
http_signature_compat,
});
let retry_sender_task = tokio::spawn(async move {
let mut join_set = JoinSet::new();
while let Some(message) = retry_receiver.recv().await {
let retry_task = retry_worker(
retry_client.clone(),
timeout,
message,
let retry_task = retry_task(
retry_stats.clone(),
retry_worker.clone(),
message,
retry_strategy,
http_signature_compat,
);
if retry_count > 0 {
@ -145,14 +168,11 @@ impl SimpleQueue {
let mut join_set = JoinSet::new();
while let Some(message) = receiver.recv().await {
let task = activity_worker(
client.clone(),
timeout,
let task = main_task(
sender_stats.clone(),
worker.clone(),
message,
retry_sender.clone(),
sender_stats.clone(),
strategy,
http_signature_compat,
);
if worker_count > 0 {
@ -180,15 +200,12 @@ impl SimpleQueue {
sender,
sender_task,
retry_sender_task,
worker_count,
}
}
#[allow(unused)]
// Drops all the senders and shuts down the workers
pub(crate) async fn shutdown(
self,
wait_for_retries: bool,
) -> Result<Arc<Stats>, anyhow::Error> {
/// Drops all the senders and shuts down the workers
pub async fn shutdown(self, wait_for_retries: bool) -> Result<Arc<Stats>, anyhow::Error> {
drop(self.sender);
self.sender_task.await?;
@ -200,3 +217,54 @@ impl SimpleQueue {
Ok(self.stats)
}
}
pub(super) async fn main_task(
stats: Arc<Stats>,
worker: Arc<Worker>,
message: ActivityMessage,
retry_queue: UnboundedSender<ActivityMessage>,
) {
stats.pending.fetch_sub(1, Ordering::Relaxed);
stats.running.fetch_add(1, Ordering::Relaxed);
let outcome = worker.queue(message.clone()).await;
// "Running" has finished, check the outcome
stats.running.fetch_sub(1, Ordering::Relaxed);
match outcome {
Ok(_) => {
stats.completed_last_hour.fetch_add(1, Ordering::Relaxed);
}
Err(_err) => {
stats.retries.fetch_add(1, Ordering::Relaxed);
warn!(
"Sending activity {} to {} to the retry queue to be tried again later",
message.activity_id, message.inbox
);
// Send to the retry queue. Ignoring whether it succeeds or not
retry_queue.send(message).ok();
}
}
}
pub(super) async fn retry_task(
stats: Arc<Stats>,
worker: Arc<Worker>,
message: ActivityMessage,
strategy: RetryStrategy,
) {
// Because the times are pretty extravagant between retries, we have to re-sign each time
let outcome = retry(|| worker.queue(message.clone()), strategy).await;
stats.retries.fetch_sub(1, Ordering::Relaxed);
match outcome {
Ok(_) => {
stats.completed_last_hour.fetch_add(1, Ordering::Relaxed);
}
Err(_err) => {
stats.dead_last_hour.fetch_add(1, Ordering::Relaxed);
}
}
}

13
src/queue/util.rs
View file

@ -1,3 +1,4 @@
//! Utilities for working with futures
use futures_core::Future;
use std::{
fmt::{Debug, Display},
@ -6,8 +7,9 @@ use std::{
use tracing::warn;
#[derive(Clone, Copy, Default)]
pub(crate) struct RetryStrategy {
/// Amount of time in seconds to back off
/// A strategy for retrying a future
pub struct RetryStrategy {
/// Amount of time in seconds to back off, exponential based upon the amount of retries i.e, sleep is backoff ^ retries
pub backoff: usize,
/// Amount of times to retry
pub retries: usize,
@ -18,12 +20,7 @@ pub(crate) struct RetryStrategy {
}
/// Retries a future action factory function up to `amount` times with an exponential backoff timer between tries
pub(crate) async fn retry<
T,
E: Display + Debug,
F: Future<Output = Result<T, E>>,
A: FnMut() -> F,
>(
pub async fn retry<T, E: Display + Debug, F: Future<Output = Result<T, E>>, A: FnMut() -> F>(
mut action: A,
strategy: RetryStrategy,
) -> Result<T, E> {

127
src/queue/worker.rs
View file

@ -1,96 +1,55 @@
//! Worker for sending activity messages
use anyhow::Error;
use async_trait::async_trait;
use derive_builder::Builder;
use reqwest_middleware::ClientWithMiddleware;
use std::{
sync::{atomic::Ordering, Arc},
time::Duration,
};
use tokio::sync::mpsc::UnboundedSender;
use tracing::warn;
use std::time::Duration;
use super::{
request::sign_and_send,
util::{retry, RetryStrategy},
SendActivityTask,
Stats,
};
use crate::config::FederationConfig;
/// A tokio spawned worker which is responsible for submitting requests to federated servers
/// This will retry up to one time with the same signature, and if it fails, will move it to the retry queue.
/// We need to retry activity sending in case the target instances is temporarily unreachable.
/// In this case, the task is stored and resent when the instance is hopefully back up. This
/// list shows the retry intervals, and which events of the target instance can be covered:
/// - 60s (one minute, service restart) -- happens in the worker w/ same signature
/// - 60min (one hour, instance maintenance) --- happens in the retry worker
/// - 60h (2.5 days, major incident with rebuild from backup) --- happens in the retry worker
pub(super) async fn activity_worker(
client: ClientWithMiddleware,
timeout: Duration,
message: SendActivityTask,
retry_queue: UnboundedSender<SendActivityTask>,
stats: Arc<Stats>,
strategy: RetryStrategy,
http_signature_compat: bool,
) {
stats.pending.fetch_sub(1, Ordering::Relaxed);
stats.running.fetch_add(1, Ordering::Relaxed);
use super::{request::sign_and_send, util::RetryStrategy, ActivityMessage, ActivityQueue};
/// Configuration for the worker, with various tweaks
#[derive(Builder, Clone)]
pub struct Worker {
#[builder(default = "reqwest::Client::default().into()")]
/// The Reqwest client to make requests with
pub client: ClientWithMiddleware,
#[builder(default = "Duration::from_secs(10)")]
/// The timeout before a request fails
pub request_timeout: Duration,
#[builder(default)]
/// The strategy for retrying requests
pub strategy: RetryStrategy,
#[builder(default = "false")]
/// Whether to enable signature compat or not
pub http_signature_compat: bool,
}
let outcome = sign_and_send(&message, &client, timeout, strategy, http_signature_compat).await;
// "Running" has finished, check the outcome
stats.running.fetch_sub(1, Ordering::Relaxed);
match outcome {
Ok(_) => {
stats.completed_last_hour.fetch_add(1, Ordering::Relaxed);
}
Err(_err) => {
stats.retries.fetch_add(1, Ordering::Relaxed);
warn!(
"Sending activity {} to {} to the retry queue to be tried again later",
message.activity_id, message.inbox
);
// Send to the retry queue. Ignoring whether it succeeds or not
retry_queue.send(message).ok();
impl Worker {
/// Create a worker from a config
pub fn from_config<T: Clone>(config: &FederationConfig<T>) -> Self {
Self {
client: config.client.clone(),
request_timeout: config.request_timeout,
strategy: Default::default(),
http_signature_compat: config.http_signature_compat,
}
}
}
pub(super) async fn retry_worker(
client: ClientWithMiddleware,
timeout: Duration,
message: SendActivityTask,
stats: Arc<Stats>,
strategy: RetryStrategy,
http_signature_compat: bool,
) {
// Because the times are pretty extravagant between retries, we have to re-sign each time
let outcome = retry(
|| {
sign_and_send(
&message,
&client,
timeout,
RetryStrategy {
backoff: 0,
retries: 0,
offset: 0,
initial_sleep: 0,
},
http_signature_compat,
)
},
strategy,
)
.await;
#[async_trait]
impl ActivityQueue for Worker {
type Error = anyhow::Error;
stats.retries.fetch_sub(1, Ordering::Relaxed);
match outcome {
Ok(_) => {
stats.completed_last_hour.fetch_add(1, Ordering::Relaxed);
}
Err(_err) => {
stats.dead_last_hour.fetch_add(1, Ordering::Relaxed);
}
async fn queue(&self, message: ActivityMessage) -> Result<(), Error> {
sign_and_send(
&message,
&self.client,
self.request_timeout,
self.strategy,
self.http_signature_compat,
)
.await
}
}