mirror of
https://github.com/openai/codex.git
synced 2026-11-18 22:20:48 +00:00
Warn users when SQLite diagnostic log writes fail (#47887)
## Why Failed SQLite writes leave diagnostic logs incomplete. Notify users before they close Codex so they can submit logs or run `codex doctor`. ## What changed - Add a best-effort warning on the first log write failure in standalone and in-process app servers, while continuing to retain failure diagnostics in memory. - Handle failures that occurred before the in-process warning reporter was attached. - Tailor the warning's guidance to whether submitting logs is enabled. ## Testing Add coverage for warning deduplication, both guidance variants, retained diagnostics after the notification sender is dropped, and SQLite flush failures before and after reporter attachment. Extend log upload integration coverage and add a TUI warnings snapshot. GitOrigin-RevId: 12a74126d13c3568430e539a180cca47fa931fa9
This commit is contained in:
committed by
copyberry
parent
a285066737
commit
3d6428510f
@@ -24,6 +24,7 @@
|
||||
use std::future::Future;
|
||||
use std::sync::Arc;
|
||||
use std::sync::OnceLock;
|
||||
use std::sync::RwLock;
|
||||
use std::sync::atomic::AtomicBool;
|
||||
use std::sync::atomic::Ordering;
|
||||
use std::time::Duration;
|
||||
@@ -136,6 +137,7 @@ pub trait LogWriteFailureReporter: Send + Sync {
|
||||
pub struct LogDbLayer {
|
||||
sender: mpsc::Sender<LogDbCommand>,
|
||||
has_write_failure: Arc<AtomicBool>,
|
||||
failure_reporter: Arc<RwLock<Arc<dyn LogWriteFailureReporter>>>,
|
||||
process_uuid: String,
|
||||
}
|
||||
|
||||
@@ -152,6 +154,7 @@ impl Clone for LogDbLayer {
|
||||
Self {
|
||||
sender: self.sender.clone(),
|
||||
has_write_failure: self.has_write_failure.clone(),
|
||||
failure_reporter: self.failure_reporter.clone(),
|
||||
process_uuid: self.process_uuid.clone(),
|
||||
}
|
||||
}
|
||||
@@ -173,20 +176,30 @@ impl LogDbLayer {
|
||||
let config = config.normalized();
|
||||
let (sender, receiver) = mpsc::channel(config.queue_capacity);
|
||||
let has_write_failure = Arc::new(AtomicBool::new(false));
|
||||
let failure_reporter = Arc::new(RwLock::new(failure_reporter));
|
||||
tokio::spawn(run_inserter(
|
||||
state_db,
|
||||
receiver,
|
||||
config,
|
||||
failure_reporter,
|
||||
failure_reporter.clone(),
|
||||
has_write_failure.clone(),
|
||||
));
|
||||
Self {
|
||||
sender,
|
||||
has_write_failure,
|
||||
failure_reporter,
|
||||
process_uuid: current_process_log_uuid().to_string(),
|
||||
}
|
||||
}
|
||||
|
||||
/// Replaces the startup reporter once the client notification destination is available.
|
||||
pub fn set_failure_reporter(&self, reporter: Arc<dyn LogWriteFailureReporter>) {
|
||||
*self
|
||||
.failure_reporter
|
||||
.write()
|
||||
.unwrap_or_else(std::sync::PoisonError::into_inner) = reporter;
|
||||
}
|
||||
|
||||
/// Whether this writer has lost logs, making SQLite incomplete for feedback.
|
||||
pub fn has_write_failure(&self) -> bool {
|
||||
self.has_write_failure.load(Ordering::Relaxed)
|
||||
@@ -479,7 +492,7 @@ async fn run_inserter(
|
||||
state_db: std::sync::Arc<StateRuntime>,
|
||||
mut receiver: mpsc::Receiver<LogDbCommand>,
|
||||
config: LogSinkQueueConfig,
|
||||
failure_reporter: Arc<dyn LogWriteFailureReporter>,
|
||||
failure_reporter: Arc<RwLock<Arc<dyn LogWriteFailureReporter>>>,
|
||||
has_write_failure: Arc<AtomicBool>,
|
||||
) {
|
||||
let mut buffer = Vec::with_capacity(config.batch_size);
|
||||
@@ -516,7 +529,7 @@ async fn run_inserter(
|
||||
async fn flush(
|
||||
state_db: &StateRuntime,
|
||||
buffer: &mut Vec<LogEntry>,
|
||||
failure_reporter: &dyn LogWriteFailureReporter,
|
||||
failure_reporter: &RwLock<Arc<dyn LogWriteFailureReporter>>,
|
||||
has_write_failure: &AtomicBool,
|
||||
) {
|
||||
if buffer.is_empty() {
|
||||
@@ -534,10 +547,10 @@ async fn flush(
|
||||
let timestamp =
|
||||
chrono::Utc::now().to_rfc3339_opts(chrono::SecondsFormat::Millis, /*use_z*/ true);
|
||||
let entry_count = entries.len();
|
||||
failure_reporter.report_failure(&format!(
|
||||
has_write_failure.store(true, Ordering::Relaxed);
|
||||
failure_reporter.read().unwrap_or_else(std::sync::PoisonError::into_inner).report_failure(&format!(
|
||||
"{timestamp} ERROR failed to flush logs to SQLite error={error:?} entries={entry_count}\n"
|
||||
));
|
||||
has_write_failure.store(true, Ordering::Relaxed);
|
||||
}
|
||||
|
||||
crate::telemetry::record_log_write(
|
||||
@@ -896,6 +909,7 @@ mod tests {
|
||||
let layer = LogDbLayer {
|
||||
sender,
|
||||
has_write_failure: Arc::new(AtomicBool::new(false)),
|
||||
failure_reporter: Arc::new(RwLock::new(Arc::new(SharedWriter::default()))),
|
||||
process_uuid: "process-1".to_string(),
|
||||
};
|
||||
|
||||
@@ -917,6 +931,7 @@ mod tests {
|
||||
let layer = LogDbLayer {
|
||||
sender,
|
||||
has_write_failure: Arc::new(AtomicBool::new(false)),
|
||||
failure_reporter: Arc::new(RwLock::new(Arc::new(SharedWriter::default()))),
|
||||
process_uuid: "process-1".to_string(),
|
||||
};
|
||||
|
||||
|
||||
Reference in New Issue
Block a user