Files
rustfs/crates/obs/src/telemetry/local.rs
T
2026-07-08 10:42:09 +00:00

904 lines
37 KiB
Rust

// Copyright 2024 RustFS Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//! Local logging backend: stdout-only or file-rolling with optional stdout mirror.
//!
//! # Behaviour
//!
//! | Condition | Result |
//! |----------------------------------|----------------------------------------------|
//! | No log directory configured | JSON logs written to **stdout only** |
//! | Log directory configured | JSON logs written to **rolling file**; |
//! | | stdout mirror enabled when `log_stdout_enabled` |
//! | | is `true` or environment is non-production |
//!
//! The function [`init_local_logging`] is the single entry point for both
//! cases; callers do **not** need to distinguish between stdout and file modes.
//!
//! The file-backed mode delegates retention and compression to
//! [`crate::cleaner`], which keeps the logging setup code focused on subscriber
//! construction while still allowing periodic housekeeping in the background.
use super::guard::OtelGuard;
use crate::TelemetryError;
use crate::cleaner::LogCleaner;
use crate::cleaner::types::{CompressionAlgorithm, FileMatchMode};
use crate::config::OtelConfig;
use crate::global::{METRIC_LOG_CLEANER_RUN_FAILURES_TOTAL, METRIC_LOG_CLEANER_RUNS_TOTAL, set_observability_metric_enabled};
use crate::telemetry::filter::build_env_filter;
use crate::telemetry::rolling::{RollingAppender, Rotation};
use metrics::counter;
use rustfs_config::observability::{
DEFAULT_OBS_LOG_CLEANUP_INTERVAL_SECONDS, DEFAULT_OBS_LOG_COMPRESS_OLD_FILES, DEFAULT_OBS_LOG_COMPRESSED_FILE_RETENTION_DAYS,
DEFAULT_OBS_LOG_COMPRESSION_ALGORITHM, DEFAULT_OBS_LOG_DELETE_EMPTY_FILES, DEFAULT_OBS_LOG_DRY_RUN,
DEFAULT_OBS_LOG_GZIP_COMPRESSION_LEVEL, DEFAULT_OBS_LOG_MATCH_MODE, DEFAULT_OBS_LOG_MAX_SINGLE_FILE_SIZE_BYTES,
DEFAULT_OBS_LOG_MAX_TOTAL_SIZE_BYTES, DEFAULT_OBS_LOG_MIN_FILE_AGE_SECONDS, DEFAULT_OBS_LOG_PARALLEL_COMPRESS,
DEFAULT_OBS_LOG_PARALLEL_WORKERS, DEFAULT_OBS_LOG_ZSTD_COMPRESSION_LEVEL, DEFAULT_OBS_LOG_ZSTD_FALLBACK_TO_GZIP,
DEFAULT_OBS_LOG_ZSTD_WORKERS,
};
use rustfs_config::{APP_NAME, DEFAULT_LOG_KEEP_FILES, DEFAULT_LOG_ROTATION_TIME, DEFAULT_OBS_LOG_STDOUT_ENABLED};
use serde_json::Value as JsonValue;
use std::sync::Arc;
use std::{collections::BTreeMap, fmt};
use std::{fs, io::IsTerminal, time::Duration};
use tracing::Subscriber;
use tracing::{info, warn};
use tracing_error::ErrorLayer;
use tracing_subscriber::{
fmt::{
FormatEvent,
format::{FmtSpan, Format, Json, JsonFields, Writer},
time::LocalTime,
},
layer::SubscriberExt,
registry::LookupSpan,
util::SubscriberInitExt,
};
const LOG_COMPONENT_OBS: &str = "obs";
const LOG_SUBSYSTEM_LOCAL_LOGGING: &str = "local_logging";
const EVENT_LOCAL_LOGGING_STATE: &str = "local_logging_state";
const EVENT_LOG_CLEANER_STATE: &str = "log_cleaner_state";
const STDERR_WARNING_PREFIX: &str = "[WARN]";
const REQUEST_ID_CANONICAL: &str = "request_id";
const REQUEST_ID_COMPAT: &str = "request-id";
fn resolve_log_cleanup_interval_seconds(config: &OtelConfig) -> u64 {
match config.log_cleanup_interval_seconds {
Some(0) => {
warn!(
event = EVENT_LOG_CLEANER_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
result = "invalid_cleanup_interval",
configured_seconds = 0_u64,
fallback_seconds = DEFAULT_OBS_LOG_CLEANUP_INTERVAL_SECONDS,
"log cleaner state changed"
);
DEFAULT_OBS_LOG_CLEANUP_INTERVAL_SECONDS
}
Some(seconds) => seconds,
None => DEFAULT_OBS_LOG_CLEANUP_INTERVAL_SECONDS,
}
}
#[derive(Clone, Debug)]
struct RequestIdJsonFormat<T> {
inner: Format<Json, T>,
}
impl<T> RequestIdJsonFormat<T> {
fn new(inner: Format<Json, T>) -> Self {
Self { inner }
}
}
impl<S, T> FormatEvent<S, JsonFields> for RequestIdJsonFormat<T>
where
S: Subscriber + for<'lookup> LookupSpan<'lookup>,
T: tracing_subscriber::fmt::time::FormatTime,
{
fn format_event(
&self,
ctx: &tracing_subscriber::fmt::FmtContext<'_, S, JsonFields>,
mut writer: Writer<'_>,
event: &tracing::Event<'_>,
) -> fmt::Result {
let mut buffer = String::new();
self.inner.format_event(ctx, Writer::new(&mut buffer), event)?;
let request_id = span_scope_request_id(ctx, event);
if request_id.is_none() {
return writer.write_str(&buffer);
}
let trimmed = buffer.trim_end();
let mut payload: JsonValue = serde_json::from_str(trimmed).map_err(|_| fmt::Error)?;
if let Some(object) = payload.as_object_mut() {
let request_id = request_id.expect("checked is_some");
object
.entry(REQUEST_ID_CANONICAL.to_string())
.or_insert_with(|| JsonValue::String(request_id.clone()));
// Keep a top-level compatibility alias for operators or downstream
// queries that already rely on the earlier hyphenated field name.
object
.entry(REQUEST_ID_COMPAT.to_string())
.or_insert_with(|| JsonValue::String(request_id));
}
let serialized = serde_json::to_string(&payload).map_err(|_| fmt::Error)?;
writer.write_str(&serialized)?;
writer.write_char('\n')
}
}
fn span_scope_request_id<S>(
ctx: &tracing_subscriber::fmt::FmtContext<'_, S, JsonFields>,
event: &tracing::Event<'_>,
) -> Option<String>
where
S: Subscriber + for<'lookup> LookupSpan<'lookup>,
{
let current_span = event.parent().and_then(|id| ctx.span(id)).or_else(|| ctx.lookup_current())?;
let mut request_id = None;
for span in current_span.scope().from_root() {
let extensions = span.extensions();
let formatted_fields = extensions.get::<tracing_subscriber::fmt::FormattedFields<JsonFields>>()?;
let fields: BTreeMap<String, JsonValue> = serde_json::from_str(&formatted_fields.fields).ok()?;
if let Some(value) = fields
.get(REQUEST_ID_CANONICAL)
.or_else(|| fields.get(REQUEST_ID_COMPAT))
.and_then(JsonValue::as_str)
{
request_id = Some(value.to_string());
}
}
request_id
}
pub(super) fn build_json_log_layer<S, W>(writer: W, enable_ansi: bool, span_events: FmtSpan) -> impl tracing_subscriber::Layer<S>
where
S: Subscriber + for<'span> LookupSpan<'span>,
W: for<'writer> tracing_subscriber::fmt::MakeWriter<'writer> + Send + Sync + 'static,
{
tracing_subscriber::fmt::layer()
.with_timer(LocalTime::rfc_3339())
.with_target(true)
.with_ansi(enable_ansi)
.with_thread_names(true)
.with_thread_ids(true)
.with_file(true)
.with_line_number(true)
.with_writer(writer)
.json()
.flatten_event(true)
.with_current_span(true)
.with_span_list(true)
.with_span_events(span_events)
.map_event_format(RequestIdJsonFormat::new)
}
/// Initialize local logging (stdout-only or file-rolling).
///
/// When `log_directory` is empty or `None` in the config the function sets up
/// a non-blocking JSON subscriber that writes to **stdout** and returns
/// immediately — no file I/O, no cleanup task.
///
/// When a log directory is provided the function additionally:
/// 1. Creates the directory (including on Unix, enforces `0755` permissions).
/// 2. Attaches a rolling-file appender (daily or hourly based on
/// `log_rotation_time`).
/// 3. Optionally mirrors output to stdout based on `log_stdout_enabled`.
/// 4. Spawns a background cleanup task that periodically removes or compresses
/// old log files according to the cleanup configuration in [`OtelConfig`].
///
/// # Arguments
/// * `config` - Observability configuration, fully populated from environment variables.
/// * `logger_level` - Effective log level string (e.g., `"info"`).
/// * `is_production` - Whether the runtime environment is production; controls
/// span verbosity and stdout mirroring defaults.
///
/// # Returns
/// An [`OtelGuard`] that keeps the `tracing_appender` worker alive and holds
/// a handle to the cleanup task (if started). Dropping the guard flushes
/// in-flight logs and stops the cleanup task.
///
/// # Errors
/// Returns [`TelemetryError`] if the log directory cannot be created or its
/// permissions cannot be set (Unix only).
pub(super) fn init_local_logging(
config: &OtelConfig,
logger_level: &str,
is_production: bool,
) -> Result<OtelGuard, TelemetryError> {
// Determine the effective log directory. An absent or empty value means
// stdout-only mode: we skip file setup entirely.
let log_dir_str = config.log_directory.as_deref().filter(|s| !s.is_empty());
if let Some(log_directory) = log_dir_str {
match init_file_logging_internal(config, log_directory, logger_level, is_production) {
Ok(guard) => Ok(guard),
Err(error) if should_fallback_to_stdout(&error) => {
let guard = init_stdout_only(config, logger_level, is_production)?;
emit_file_logging_fallback_warning(log_directory, &error);
Ok(guard)
}
Err(error) => Err(error),
}
} else {
init_stdout_only(config, logger_level, is_production)
}
}
// ─── Stdout-only ─────────────────────────────────────────────────────────────
/// Set up a non-blocking stdout JSON subscriber with no file I/O.
///
/// Used when no log directory has been configured. The subscriber formats
/// every log record as a JSON line, including RFC-3339 timestamps, thread
/// identifiers, file/line information, and span context.
///
/// # Arguments
/// * `_config` - Unused at the moment; reserved for future configuration.
/// * `logger_level` - Effective log level string.
/// * `is_production` - Controls span event verbosity.
fn init_stdout_only(_config: &OtelConfig, logger_level: &str, is_production: bool) -> Result<OtelGuard, TelemetryError> {
let env_filter = build_env_filter(logger_level, None);
let (nb, guard) = tracing_appender::non_blocking(std::io::stdout());
let span_events = if is_production { FmtSpan::CLOSE } else { FmtSpan::FULL };
// Keep stdout formatting JSON-shaped even in local-only mode so operators
// can ship the same log schema to external collectors if needed.
let fmt_layer = build_json_log_layer(nb, std::io::stdout().is_terminal(), span_events);
tracing_subscriber::registry()
.with(env_filter)
.with(ErrorLayer::default())
.with(fmt_layer)
.try_init()
.map_err(|err| TelemetryError::SubscriberInit(err.to_string()))?;
set_observability_metric_enabled(false);
counter!("rustfs_start_total").increment(1);
info!(
event = EVENT_LOCAL_LOGGING_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
state = "initialized",
backend = "local",
sink = "stdout",
output_format = "json",
logger_level,
is_production,
"local logging state"
);
Ok(OtelGuard {
tracer_provider: None,
meter_provider: None,
logger_provider: None,
profiling_agent: None,
tracing_guard: Some(guard),
stdout_guard: None,
cleanup_handle: None,
})
}
// ─── File-rolling ─────────────────────────────────────────────────────────────
/// Internal implementation for file-based rolling log setup.
///
/// Called by [`init_local_logging`] when a log directory is present.
/// Handles directory creation, permission enforcement (Unix), file appender
/// setup, optional stdout mirror, and log-cleanup task spawning.
///
/// The function intentionally performs all fallible filesystem preparation
/// before registering the subscriber so startup failures are reported early and
/// do not leave partially initialized tracing state behind.
fn init_file_logging_internal(
config: &OtelConfig,
log_directory: &str,
logger_level: &str,
is_production: bool,
) -> Result<OtelGuard, TelemetryError> {
let service_name = config.service_name.as_deref().unwrap_or(APP_NAME);
let log_filename = config.log_filename.as_deref().unwrap_or(service_name);
let keep_files = config.log_keep_files.unwrap_or(DEFAULT_LOG_KEEP_FILES);
// ── 1. Ensure the log directory exists ───────────────────────────────────
if let Err(e) = fs::create_dir_all(log_directory) {
return Err(TelemetryError::Io(e.to_string()));
}
// ── 2. Enforce directory permissions (Unix only) ─────────────────────────
#[cfg(unix)]
ensure_dir_permissions(log_directory)?;
// ── 3. Choose rotation strategy ──────────────────────────────────────────
// `log_rotation_time` drives the rolling-appender rotation period.
let rotation_str = config
.log_rotation_time
.as_deref()
.unwrap_or(DEFAULT_LOG_ROTATION_TIME)
.to_lowercase();
// Match mode controls how the rolling filename is recognized later by the
// cleaner. Suffix mode fits timestamp-prefixed filenames especially well.
let match_mode = FileMatchMode::from_config_str(config.log_match_mode.as_deref().unwrap_or(DEFAULT_OBS_LOG_MATCH_MODE));
let rotation = match rotation_str.as_str() {
"minutely" => Rotation::Minutely,
"hourly" => Rotation::Hourly,
"daily" => Rotation::Daily,
_ => Rotation::Daily,
};
let max_single_file_size = config
.log_max_single_file_size_bytes
.unwrap_or(DEFAULT_OBS_LOG_MAX_SINGLE_FILE_SIZE_BYTES);
let file_appender =
RollingAppender::new(log_directory, log_filename.to_string(), rotation, max_single_file_size, match_mode)?;
let (non_blocking, guard) = tracing_appender::non_blocking(file_appender);
// ── 4. Build subscriber layers ────────────────────────────────────────────
let env_filter = build_env_filter(logger_level, None);
let span_events = if is_production { FmtSpan::CLOSE } else { FmtSpan::FULL };
// File output stays machine-readable and free of ANSI sequences so the
// resulting files are safe to parse or ship to log processors.
let file_layer = build_json_log_layer(non_blocking, false, span_events.clone());
// Optional stdout mirror: enabled explicitly via `log_stdout_enabled`, or
// unconditionally in non-production environments so developers still see
// immediate terminal output while file rotation remains enabled.
let (stdout_layer, stdout_guard) = if config.log_stdout_enabled.unwrap_or(DEFAULT_OBS_LOG_STDOUT_ENABLED) || !is_production {
let (stdout_nb, stdout_guard) = tracing_appender::non_blocking(std::io::stdout());
let enable_color = std::io::stdout().is_terminal();
(Some(build_json_log_layer(stdout_nb, enable_color, span_events)), Some(stdout_guard))
} else {
(None, None)
};
tracing_subscriber::registry()
.with(env_filter)
.with(ErrorLayer::default())
.with(file_layer)
.with(stdout_layer)
.try_init()
.map_err(|err| TelemetryError::SubscriberInit(err.to_string()))?;
set_observability_metric_enabled(false);
// ── 5. Start background cleanup task ─────────────────────────────────────
let cleanup_handle = spawn_cleanup_task(config, log_directory, log_filename, keep_files);
info!(
event = EVENT_LOCAL_LOGGING_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
state = "initialized",
backend = "local",
sink = "file",
output_format = "json",
log_directory,
rotation = %rotation_str,
keep_files,
stdout_mirror_enabled = stdout_guard.is_some(),
is_production,
logger_level,
"local logging state"
);
Ok(OtelGuard {
tracer_provider: None,
meter_provider: None,
logger_provider: None,
profiling_agent: None,
tracing_guard: Some(guard),
stdout_guard,
cleanup_handle: Some(cleanup_handle),
})
}
// ─── Directory permissions (Unix) ─────────────────────────────────────────────
/// Ensure the log directory has at most `0755` permissions (Unix only).
///
/// Tightens permissions to `0755` if the directory is more permissive.
/// This prevents world-writable log directories from being a security hazard.
/// No-ops if permissions are already `0755` or stricter.
///
/// The function never broadens permissions; it is strictly a hardening step.
#[cfg(unix)]
pub fn ensure_dir_permissions(log_directory: &str) -> Result<(), TelemetryError> {
use std::fs::Permissions;
use std::os::unix::fs::PermissionsExt;
let desired: u32 = 0o755;
match fs::metadata(log_directory) {
Ok(meta) => {
let current = meta.permissions().mode() & 0o777;
// Only tighten to 0755 if existing permissions are looser than target.
if (current & !desired) != 0 {
if let Err(e) = fs::set_permissions(log_directory, Permissions::from_mode(desired)) {
return Err(TelemetryError::SetPermissions(format!(
"dir='{log_directory}', want={desired:#o}, have={current:#o}, err={e}"
)));
}
// Second verification pass to confirm the change took effect.
if let Ok(meta2) = fs::metadata(log_directory) {
let after = meta2.permissions().mode() & 0o777;
if after != desired {
return Err(TelemetryError::SetPermissions(format!(
"dir='{log_directory}', want={desired:#o}, after={after:#o}"
)));
}
}
}
Ok(())
}
Err(e) => Err(TelemetryError::Io(format!("stat '{log_directory}' failed: {e}"))),
}
}
pub(super) fn should_fallback_to_stdout(error: &TelemetryError) -> bool {
match error {
TelemetryError::SetPermissions(_) => true,
TelemetryError::Io(message) => {
let message = message.to_ascii_lowercase();
message.contains("permission denied") || message.contains("os error 13")
}
_ => false,
}
}
fn format_file_logging_fallback_warning(log_directory: &str, error: &TelemetryError) -> String {
format!(
"{STDERR_WARNING_PREFIX} local logging fallback to stdout: failed_sink=file sink=stdout log_directory={log_directory} error={error}"
)
}
pub(super) fn emit_file_logging_fallback_warning(log_directory: &str, error: &TelemetryError) {
if tracing::dispatcher::has_been_set() {
warn!(
event = EVENT_LOCAL_LOGGING_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
state = "fallback_to_stdout",
backend = "local",
failed_sink = "file",
sink = "stdout",
log_directory,
error = %error,
"local logging state changed"
);
} else {
eprintln!("{}", format_file_logging_fallback_warning(log_directory, error));
}
}
// ─── Cleanup task ─────────────────────────────────────────────────────────────
/// Spawn a background task that periodically cleans up old log files.
///
/// All cleanup parameters are derived from [`OtelConfig`] fields, with
/// sensible defaults when fields are absent. The task runs on the current
/// Tokio runtime and should be aborted (via the returned `JoinHandle`) when
/// the application shuts down.
///
/// The asynchronous loop itself remains lightweight: each cleanup pass is
/// delegated to `spawn_blocking` because directory traversal, compression, and
/// deletion are inherently blocking filesystem operations.
///
/// # Arguments
/// * `config` - Observability config containing cleanup parameters.
/// * `log_directory` - Directory path of the rolling log files.
/// * `log_filename` - Base filename (used as the file prefix for matching).
/// * `keep_files` - Legacy keep-files count; used as fallback when the new
/// `log_keep_count` field is absent.
///
/// # Returns
/// A [`tokio::task::JoinHandle`] for the spawned cleanup loop.
pub fn spawn_cleanup_task(
config: &OtelConfig,
log_directory: &str,
log_filename: &str,
keep_files: usize,
) -> tokio::task::JoinHandle<()> {
let log_dir = std::path::PathBuf::from(log_directory);
// Use suffix matching for log files like `2026-03-01-06-21.rustfs.log`
// where `rustfs.log` is the stable suffix generated by the rolling appender.
let file_pattern = config.log_filename.as_deref().unwrap_or(log_filename).to_string();
let active_filename = file_pattern.clone();
// Determine match mode from config, defaulting to the repository-wide
// observability setting when the caller leaves it unset.
let match_mode = FileMatchMode::from_config_str(config.log_match_mode.as_deref().unwrap_or(DEFAULT_OBS_LOG_MATCH_MODE));
let max_total_size = config
.log_max_total_size_bytes
.unwrap_or(DEFAULT_OBS_LOG_MAX_TOTAL_SIZE_BYTES);
let max_single_file_size = config
.log_max_single_file_size_bytes
.unwrap_or(DEFAULT_OBS_LOG_MAX_SINGLE_FILE_SIZE_BYTES);
let compress = config.log_compress_old_files.unwrap_or(DEFAULT_OBS_LOG_COMPRESS_OLD_FILES);
let gzip_level = config
.log_gzip_compression_level
.unwrap_or(DEFAULT_OBS_LOG_GZIP_COMPRESSION_LEVEL);
let compression_algorithm = CompressionAlgorithm::from_config_str(
config
.log_compression_algorithm
.as_deref()
.unwrap_or(DEFAULT_OBS_LOG_COMPRESSION_ALGORITHM),
);
let parallel_compress = config.log_parallel_compress.unwrap_or(DEFAULT_OBS_LOG_PARALLEL_COMPRESS);
let parallel_workers = config.log_parallel_workers.unwrap_or(DEFAULT_OBS_LOG_PARALLEL_WORKERS);
let zstd_level = config
.log_zstd_compression_level
.unwrap_or(DEFAULT_OBS_LOG_ZSTD_COMPRESSION_LEVEL);
let zstd_fallback_to_gzip = config
.log_zstd_fallback_to_gzip
.unwrap_or(DEFAULT_OBS_LOG_ZSTD_FALLBACK_TO_GZIP);
let zstd_workers = config.log_zstd_workers.unwrap_or(DEFAULT_OBS_LOG_ZSTD_WORKERS);
let retention_days = config
.log_compressed_file_retention_days
.unwrap_or(DEFAULT_OBS_LOG_COMPRESSED_FILE_RETENTION_DAYS);
let exclude_patterns = config
.log_exclude_patterns
.as_deref()
.map(|s| s.split(',').map(|p| p.trim().to_string()).collect())
.unwrap_or_default();
let delete_empty = config.log_delete_empty_files.unwrap_or(DEFAULT_OBS_LOG_DELETE_EMPTY_FILES);
let min_age = config
.log_min_file_age_seconds
.unwrap_or(DEFAULT_OBS_LOG_MIN_FILE_AGE_SECONDS);
let dry_run = config.log_dry_run.unwrap_or(DEFAULT_OBS_LOG_DRY_RUN);
let cleanup_interval = resolve_log_cleanup_interval_seconds(config);
let cleaner = Arc::new(
LogCleaner::builder(log_dir, file_pattern, active_filename)
.match_mode(match_mode)
.keep_files(keep_files)
.max_total_size_bytes(max_total_size)
.max_single_file_size_bytes(max_single_file_size)
.compress_old_files(compress)
.gzip_compression_level(gzip_level)
// Compression behavior stays fully config-driven, but the builder
// clamps unsafe numeric values and preserves sensible defaults.
.compression_algorithm(compression_algorithm)
.parallel_compress(parallel_compress)
.parallel_workers(parallel_workers)
.zstd_compression_level(zstd_level)
.zstd_fallback_to_gzip(zstd_fallback_to_gzip)
.zstd_workers(zstd_workers)
.compressed_file_retention_days(retention_days)
.exclude_patterns(exclude_patterns)
.delete_empty_files(delete_empty)
.min_file_age_seconds(min_age)
.dry_run(dry_run)
.build(),
);
info!(
event = EVENT_LOG_CLEANER_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
state = "configured",
compression_algorithm = %compression_algorithm,
parallel_compress,
parallel_workers,
zstd_level,
zstd_fallback_to_gzip,
zstd_workers,
"log cleaner state"
);
tokio::spawn(async move {
let mut interval = tokio::time::interval(Duration::from_secs(cleanup_interval));
loop {
// Wait for the next scheduled tick before dispatching another pass.
// The blocking filesystem work runs on a dedicated blocking thread.
interval.tick().await;
let cleaner_clone = cleaner.clone();
let result = tokio::task::spawn_blocking(move || cleaner_clone.cleanup()).await;
match result {
Ok(Ok(_)) => {
counter!(METRIC_LOG_CLEANER_RUNS_TOTAL).increment(1);
}
Ok(Err(e)) => {
counter!(METRIC_LOG_CLEANER_RUN_FAILURES_TOTAL).increment(1);
warn!(
event = EVENT_LOG_CLEANER_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
result = "cleanup_failed",
error = %e,
"log cleaner state changed"
);
}
Err(e) => {
counter!(METRIC_LOG_CLEANER_RUN_FAILURES_TOTAL).increment(1);
warn!(
event = EVENT_LOG_CLEANER_STATE,
component = LOG_COMPONENT_OBS,
subsystem = LOG_SUBSYSTEM_LOCAL_LOGGING,
result = "cleanup_task_panicked",
error = %e,
"log cleaner state changed"
);
}
}
}
})
}
#[cfg(test)]
mod tests {
use super::*;
use crate::config::OtelConfig;
use serde_json::Value;
use std::collections::BTreeSet;
use std::io;
use std::sync::{Arc, Mutex};
use tempfile::tempdir;
use tracing_subscriber::{Registry, layer::SubscriberExt};
#[derive(Clone, Default)]
struct SharedWriter {
inner: Arc<Mutex<Vec<u8>>>,
}
struct SharedWriterGuard {
inner: Arc<Mutex<Vec<u8>>>,
}
impl<'writer> tracing_subscriber::fmt::MakeWriter<'writer> for SharedWriter {
type Writer = SharedWriterGuard;
fn make_writer(&'writer self) -> Self::Writer {
SharedWriterGuard {
inner: Arc::clone(&self.inner),
}
}
}
impl io::Write for SharedWriterGuard {
fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
let mut inner = self.inner.lock().expect("shared writer lock should not be poisoned");
inner.extend_from_slice(buf);
Ok(buf.len())
}
fn flush(&mut self) -> io::Result<()> {
Ok(())
}
}
fn render_json_log(enable_ansi: bool) -> (String, Value) {
let writer = SharedWriter::default();
let layer = build_json_log_layer(writer.clone(), enable_ansi, FmtSpan::NONE);
let subscriber = Registry::default().with(layer);
tracing::subscriber::with_default(subscriber, || {
info!(component = "obs_contract_test", sink = "stdout", "hello world");
});
let raw = String::from_utf8(
writer
.inner
.lock()
.expect("shared writer lock should not be poisoned")
.clone(),
)
.expect("log output should be valid UTF-8");
let first_line = raw.lines().next().expect("expected one JSON log line");
let parsed: Value = serde_json::from_str(first_line).expect("log line should be valid JSON");
(raw, parsed)
}
fn render_json_log_with_request_span() -> Value {
let writer = SharedWriter::default();
let layer = build_json_log_layer(writer.clone(), false, FmtSpan::NONE);
let subscriber = Registry::default().with(layer);
tracing::subscriber::with_default(subscriber, || {
let span = tracing::info_span!("http-request", request_id = "req-123", method = "GET");
let _guard = span.enter();
info!(component = "obs_contract_test", message = "inside request span");
});
let raw = String::from_utf8(
writer
.inner
.lock()
.expect("shared writer lock should not be poisoned")
.clone(),
)
.expect("log output should be valid UTF-8");
let first_line = raw.lines().next().expect("expected one JSON log line");
serde_json::from_str(first_line).expect("log line should be valid JSON")
}
fn render_json_log_with_recovery_monitor_child_span() -> Value {
let writer = SharedWriter::default();
let layer = build_json_log_layer(writer.clone(), false, FmtSpan::NONE);
let subscriber = Registry::default().with(layer);
tracing::subscriber::with_default(subscriber, || {
let request_span = tracing::info_span!("request-span", request_id = "req-parent", method = "GET");
let _request_guard = request_span.enter();
let recovery_span =
tracing::info_span!("recovery-monitor", kind = "remote_disk", endpoint = "http://127.0.0.1:9000/data");
let _recovery_guard = recovery_span.enter();
warn!(component = "obs_contract_test", message = "inside recovery monitor");
});
let raw = String::from_utf8(
writer
.inner
.lock()
.expect("shared writer lock should not be poisoned")
.clone(),
)
.expect("log output should be valid UTF-8");
let first_line = raw.lines().next().expect("expected one JSON log line");
serde_json::from_str(first_line).expect("log line should be valid JSON")
}
#[test]
/// Invalid file names should be reported as errors instead of panicking.
fn test_init_file_logging_invalid_filename_does_not_panic() {
let temp_dir = tempdir().expect("create temp dir");
let temp_path = temp_dir.path().to_str().expect("temp dir path is utf-8");
let config = OtelConfig {
log_filename: Some("invalid\0name.log".to_string()),
..OtelConfig::default()
};
// We must run within a Tokio runtime because init_file_logging_internal spawns a background task.
let rt = tokio::runtime::Runtime::new().unwrap();
rt.block_on(async {
let result = init_file_logging_internal(&config, temp_path, "info", true);
// With eager file opening, an invalid filename (null byte) causes the OS to reject
// the open() call, so the function returns Err instead of panicking.
assert!(result.is_err(), "invalid filename must return Err, not panic");
});
}
#[test]
fn test_permission_denied_errors_fall_back_to_stdout() {
assert!(should_fallback_to_stdout(&TelemetryError::Io(
"Permission denied (os error 13)".to_string()
)));
assert!(should_fallback_to_stdout(&TelemetryError::SetPermissions(
"dir='/logs', want=0o755, have=0o777, err=Permission denied (os error 13)".to_string()
)));
assert!(!should_fallback_to_stdout(&TelemetryError::Io(
"No such file or directory (os error 2)".to_string()
)));
}
#[test]
fn test_file_logging_fallback_warning_message_is_actionable() {
let message = format_file_logging_fallback_warning(
"/var/log/rustfs",
&TelemetryError::Io("Permission denied (os error 13)".to_string()),
);
assert!(message.starts_with(STDERR_WARNING_PREFIX));
assert!(message.contains("fallback to stdout"));
assert!(message.contains("log_directory=/var/log/rustfs"));
assert!(message.contains("Permission denied"));
}
#[test]
fn test_json_log_layer_produces_parseable_json_without_ansi_escape_sequences() {
let (raw, parsed) = render_json_log(true);
assert!(!raw.contains('\u{1b}'), "JSON log output must not contain ANSI escape sequences: {raw}");
assert_eq!(parsed.get("level").and_then(Value::as_str), Some("INFO"));
assert_eq!(parsed["message"], Value::String("hello world".to_string()));
assert_eq!(parsed["component"], Value::String("obs_contract_test".to_string()));
assert_eq!(parsed["sink"], Value::String("stdout".to_string()));
assert!(parsed.get("target").is_some(), "expected target field in JSON log output");
}
#[test]
fn test_json_log_layer_keeps_same_shape_when_ansi_setting_changes() {
let (_raw_plain, plain) = render_json_log(false);
let (_raw_color, color) = render_json_log(true);
let plain_keys = plain
.as_object()
.expect("plain JSON log should be an object")
.keys()
.cloned()
.collect::<BTreeSet<_>>();
let color_keys = color
.as_object()
.expect("color JSON log should be an object")
.keys()
.cloned()
.collect::<BTreeSet<_>>();
assert_eq!(plain_keys, color_keys, "top-level JSON log keys must stay stable across sinks");
let plain_field_keys = plain
.as_object()
.expect("plain JSON log should be an object")
.keys()
.filter(|key| {
!matches!(
key.as_str(),
"timestamp" | "level" | "target" | "filename" | "line_number" | "threadName" | "threadId"
)
})
.cloned()
.collect::<BTreeSet<_>>();
let color_field_keys = color
.as_object()
.expect("color JSON log should be an object")
.keys()
.filter(|key| {
!matches!(
key.as_str(),
"timestamp" | "level" | "target" | "filename" | "line_number" | "threadName" | "threadId"
)
})
.cloned()
.collect::<BTreeSet<_>>();
assert_eq!(plain_field_keys, color_field_keys, "structured field keys must stay stable across sinks");
}
#[test]
fn test_json_log_layer_promotes_request_id_from_current_span() {
let parsed = render_json_log_with_request_span();
assert_eq!(parsed["request_id"], Value::String("req-123".to_string()));
assert_eq!(parsed["request-id"], Value::String("req-123".to_string()));
assert_eq!(parsed["message"], Value::String("inside request span".to_string()));
assert_eq!(parsed["span"]["request_id"], Value::String("req-123".to_string()));
}
#[test]
fn test_json_log_layer_promotes_parent_request_id_for_recovery_monitor_child_span() {
let parsed = render_json_log_with_recovery_monitor_child_span();
assert_eq!(parsed["request_id"], Value::String("req-parent".to_string()));
assert_eq!(parsed["request-id"], Value::String("req-parent".to_string()));
assert_eq!(parsed["message"], Value::String("inside recovery monitor".to_string()));
assert_eq!(parsed["span"]["name"], Value::String("recovery-monitor".to_string()));
assert_eq!(parsed["span"]["kind"], Value::String("remote_disk".to_string()));
let spans = parsed["spans"].as_array().expect("spans should be present");
assert!(spans.iter().any(|span| {
span.get("name").and_then(Value::as_str) == Some("request-span")
&& span.get("request_id").and_then(Value::as_str) == Some("req-parent")
}));
}
#[test]
fn test_resolve_log_cleanup_interval_seconds_rejects_zero() {
let config = OtelConfig {
log_cleanup_interval_seconds: Some(0),
..OtelConfig::default()
};
assert_eq!(resolve_log_cleanup_interval_seconds(&config), DEFAULT_OBS_LOG_CLEANUP_INTERVAL_SECONDS);
}
}