Alert on frequent network errors (#7410)

* Introduce is_frequent util

* Add dirty warn_if_frequent! implementation

* Add freq

* Fix order in condition

* Update

* Update docs

* Fix

* Remove old impl

* Fix errors

* Add wif to av-distr

* Add wif to col prot

* Rename

* Add wif to state-distr

* Address review comments

* Change Freq implementation

* Remove the zero division check

* Make rate explicit

* Fix typo

* Update rate constant

* Introduce explicit rates

* Update docs

* Split errors freq

* Downgrade coarsetime
This commit is contained in:
Andrei Eres
2023-07-17 14:05:57 +02:00
committed by GitHub
parent dd7d2f924b
commit 174f23d1cc
12 changed files with 268 additions and 15 deletions
@@ -91,7 +91,11 @@ pub type Result<T> = std::result::Result<T, Error>;
///
/// We basically always want to try and continue on error. This utility function is meant to
/// consume top-level errors by simply logging them
pub fn log_error(result: Result<()>, ctx: &'static str) -> std::result::Result<(), FatalError> {
pub fn log_error(
result: Result<()>,
ctx: &'static str,
warn_freq: &mut gum::Freq,
) -> std::result::Result<(), FatalError> {
match result.into_nested()? {
Ok(()) => Ok(()),
Err(jfyi) => {
@@ -104,7 +108,8 @@ pub fn log_error(result: Result<()>, ctx: &'static str) -> std::result::Result<(
JfyiError::FetchPoV(_) |
JfyiError::SendResponse |
JfyiError::NoSuchPoV |
JfyiError::Runtime(_) => gum::debug!(target: LOG_TARGET, error = ?jfyi, ctx),
JfyiError::Runtime(_) =>
gum::warn_if_frequent!(freq: warn_freq, max_rate: gum::Times::PerHour(100), target: LOG_TARGET, error = ?jfyi, ctx),
}
Ok(())
},
@@ -97,6 +97,7 @@ impl AvailabilityDistributionSubsystem {
let IncomingRequestReceivers { pov_req_receiver, chunk_req_receiver } = recvs;
let mut requester = Requester::new(metrics.clone()).fuse();
let mut warn_freq = gum::Freq::new();
{
let sender = ctx.sender().clone();
@@ -147,6 +148,7 @@ impl AvailabilityDistributionSubsystem {
.update_fetching_heads(&mut ctx, &mut runtime, update, &spans)
.await,
"Error in Requester::update_fetching_heads",
&mut warn_freq,
)?;
},
FromOrchestra::Signal(OverseerSignal::BlockFinalized(hash, _)) => {
@@ -188,6 +190,7 @@ impl AvailabilityDistributionSubsystem {
)
.await,
"pov_requester::fetch_pov",
&mut warn_freq,
)?;
},
}
@@ -260,6 +260,8 @@ impl RunningTask {
let mut succeeded = false;
let mut count: u32 = 0;
let mut span = self.span.child("run-fetch-chunk-task").with_relay_parent(self.relay_parent);
let mut network_error_freq = gum::Freq::new();
let mut canceled_freq = gum::Freq::new();
// Try validators in reverse order:
while let Some(validator) = self.group.pop() {
// Report retries:
@@ -272,7 +274,10 @@ impl RunningTask {
.with_chunk_index(self.request.index.0)
.with_stage(jaeger::Stage::AvailabilityDistribution);
// Send request:
let resp = match self.do_request(&validator).await {
let resp = match self
.do_request(&validator, &mut network_error_freq, &mut canceled_freq)
.await
{
Ok(resp) => resp,
Err(TaskError::ShuttingDown) => {
gum::info!(
@@ -342,6 +347,8 @@ impl RunningTask {
async fn do_request(
&mut self,
validator: &AuthorityDiscoveryId,
nerwork_error_freq: &mut gum::Freq,
canceled_freq: &mut gum::Freq,
) -> std::result::Result<ChunkFetchingResponse, TaskError> {
gum::trace!(
target: LOG_TARGET,
@@ -386,7 +393,9 @@ impl RunningTask {
Err(TaskError::PeerError)
},
Err(RequestError::NetworkError(err)) => {
gum::debug!(
gum::warn_if_frequent!(
freq: nerwork_error_freq,
max_rate: gum::Times::PerHour(100),
target: LOG_TARGET,
origin = ?validator,
relay_parent = ?self.relay_parent,
@@ -400,7 +409,9 @@ impl RunningTask {
Err(TaskError::PeerError)
},
Err(RequestError::Canceled(oneshot::Canceled)) => {
gum::debug!(
gum::warn_if_frequent!(
freq: canceled_freq,
max_rate: gum::Times::PerHour(100),
target: LOG_TARGET,
origin = ?validator,
relay_parent = ?self.relay_parent,