mirror of
https://github.com/LemmyNet/lemmy.git
synced 2024-11-25 22:01:19 +00:00
move stats to own file
This commit is contained in:
parent
bc64644e60
commit
76393bd964
1 changed files with 8 additions and 63 deletions
|
@ -1,25 +1,22 @@
|
|||
use crate::{
|
||||
util::{get_latest_activity_id, CancellableTask},
|
||||
worker::InstanceWorker,
|
||||
};
|
||||
use crate::{util::CancellableTask, worker::InstanceWorker};
|
||||
use activitypub_federation::config::FederationConfig;
|
||||
use chrono::{Local, Timelike};
|
||||
use lemmy_api_common::{context::LemmyContext, federate_retry_sleep_duration};
|
||||
use lemmy_api_common::context::LemmyContext;
|
||||
use lemmy_db_schema::{
|
||||
newtypes::InstanceId,
|
||||
source::{federation_queue_state::FederationQueueState, instance::Instance},
|
||||
utils::{ActualDbPool, DbPool},
|
||||
};
|
||||
use lemmy_utils::error::LemmyResult;
|
||||
use stats::receive_print_stats;
|
||||
use std::{collections::HashMap, time::Duration};
|
||||
use tokio::{
|
||||
sync::mpsc::{unbounded_channel, UnboundedReceiver, UnboundedSender},
|
||||
sync::mpsc::{unbounded_channel, UnboundedSender},
|
||||
task::JoinHandle,
|
||||
time::{interval, sleep},
|
||||
time::sleep,
|
||||
};
|
||||
use tokio_util::sync::CancellationToken;
|
||||
use tracing::{debug, error, info};
|
||||
use tracing::info;
|
||||
|
||||
mod stats;
|
||||
mod util;
|
||||
mod worker;
|
||||
|
||||
|
@ -149,65 +146,13 @@ impl SendManager {
|
|||
}
|
||||
}
|
||||
|
||||
/// every 60s, print the state for every instance. exits if the receiver is done (all senders dropped)
|
||||
async fn receive_print_stats(
|
||||
pool: ActualDbPool,
|
||||
mut receiver: UnboundedReceiver<(String, FederationQueueState)>,
|
||||
) {
|
||||
let pool = &mut DbPool::Pool(&pool);
|
||||
let mut printerval = interval(Duration::from_secs(60));
|
||||
let mut stats = HashMap::new();
|
||||
loop {
|
||||
tokio::select! {
|
||||
ele = receiver.recv() => {
|
||||
let Some((domain, ele)) = ele else {
|
||||
print_stats(pool, &stats).await;
|
||||
return;
|
||||
};
|
||||
stats.insert(domain, ele);
|
||||
},
|
||||
_ = printerval.tick() => {
|
||||
print_stats(pool, &stats).await;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
async fn print_stats(pool: &mut DbPool<'_>, stats: &HashMap<String, FederationQueueState>) {
|
||||
let last_id = get_latest_activity_id(pool).await;
|
||||
let Ok(last_id) = last_id else {
|
||||
error!("could not get last id");
|
||||
return;
|
||||
};
|
||||
// it's expected that the values are a bit out of date, everything < SAVE_STATE_EVERY should be considered up to date
|
||||
info!("Federation state as of {}:", Local::now().to_rfc3339());
|
||||
// todo: more stats (act/sec, avg http req duration)
|
||||
let mut ok_count = 0;
|
||||
let mut behind_count = 0;
|
||||
for (domain, stat) in stats {
|
||||
let behind = last_id.0 - stat.last_successful_id.map(|e| e.0).unwrap_or(0);
|
||||
if stat.fail_count > 0 {
|
||||
info!(
|
||||
"{domain}: Warning. {behind} behind, {} consecutive fails, current retry delay {:.2?}",
|
||||
stat.fail_count,
|
||||
federate_retry_sleep_duration(stat.fail_count)
|
||||
);
|
||||
} else if behind > 0 {
|
||||
debug!("{}: Ok. {} activities behind", domain, behind);
|
||||
behind_count += 1;
|
||||
} else {
|
||||
ok_count += 1;
|
||||
}
|
||||
}
|
||||
info!("{ok_count} others up to date. {behind_count} instances behind.");
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
#[allow(clippy::unwrap_used)]
|
||||
#[allow(clippy::indexing_slicing)]
|
||||
mod test {
|
||||
|
||||
use super::*;
|
||||
use tokio::time::sleep;
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_start_stop_federation_workers() -> LemmyResult<()> {
|
||||
|
|
Loading…
Reference in a new issue