add opentelemetry for debugging and testing behind otel feature flag

This commit is contained in:
Mark Sinclair
2025-08-20 18:04:48 +01:00
committed by Floriane TUERNAL SABOTINOV
parent 88c4e0ce6c
commit 20e6243738
9 changed files with 1407 additions and 10636 deletions
Generated
+1079 -1188
View File
File diff suppressed because it is too large Load Diff
+8 -3
View File
@@ -300,8 +300,11 @@ nix = "0.27.1"
notify = "5.1.0"
okapi = "0.7.0"
once_cell = "1.21.3"
opentelemetry = "0.19.0"
opentelemetry-jaeger = "0.18.0"
opentelemetry = "0.30.0"
opentelemetry-otlp = "0.30.0"
opentelemetry-semantic-conventions = "0.30.0"
opentelemetry_sdk = "0.30.0"
opentelemetry-stdout = "0.30.0"
parking_lot = "0.12.3"
pem = "0.8"
petgraph = "0.6.5"
@@ -360,8 +363,10 @@ toml = "0.8.22"
tower = "0.5.2"
tower-http = "0.5.2"
tracing = "0.1.41"
tracing-core = "0.1.33"
tracing-log = "0.2"
tracing-opentelemetry = "0.19.0"
tracing-opentelemetry = "0.31.0"
tracing-serde = "0.2.0"
tracing-subscriber = "0.3.19"
tracing-tree = "0.2.2"
tracing-indicatif = "0.3.9"
+28 -11
View File
@@ -8,40 +8,57 @@ license = { workspace = true }
repository = { workspace = true }
[dependencies]
chrono = { workspace = true, optional = true }
cfg-if = { workspace = true }
clap = { workspace = true, features = ["derive"], optional = true }
clap_complete = { workspace = true, optional = true }
clap_complete_fig = { workspace = true, optional = true }
const-str = { workspace = true }
log = { workspace = true }
opentelemetry = { workspace = true, optional = true }
opentelemetry-otlp = { workspace = true,features=["metrics", "grpc-tonic", "tls",
"tls-webpki-roots"], optional = true }
opentelemetry-semantic-conventions = { workspace = true, features = ["semconv_experimental"], optional = true }
opentelemetry-stdout = { workspace = true, features = ["trace", "metrics"], optional = true }
opentelemetry_sdk = { workspace = true, optional = true }
schemars = { workspace = true, features = ["preserve_order"], optional = true }
serde = { workspace = true, features = ["derive"] }
serde_json = { workspace = true, optional = true }
## tracing
tracing-subscriber = { workspace = true, features = ["env-filter"], optional = true }
tracing-tree = { workspace = true, optional = true }
thiserror = { workspace = true }
tracing = { workspace = true, optional = true }
opentelemetry-jaeger = { workspace = true, features = ["rt-tokio", "collector_client", "isahc_collector_client"], optional = true }
tracing-core = { workspace = true, optional = true }
tracing-opentelemetry = { workspace = true, optional = true }
tracing-serde = { workspace = true, optional = true }
tracing-subscriber = { workspace = true, features = ["env-filter", "json"], optional = true }
tracing-tree = { workspace = true, optional = true }
utoipa = { workspace = true, optional = true }
opentelemetry = { workspace = true, features = ["rt-tokio"], optional = true }
[build-dependencies]
vergen = { workspace = true, features = ["build", "git", "gitcl", "rustc", "cargo"] }
[features]
default = []
default = ["otel"] # 🚨🚨🚨 TODO: remove otel as a default feature 🚨🚨🚨
openapi = ["utoipa"]
output_format = ["serde_json", "dep:clap"]
bin_info_schema = ["schemars"]
basic_tracing = ["dep:tracing", "tracing-subscriber"]
tokio-console = ["dep:tracing", "tracing-subscriber"]
tracing = [
"basic_tracing",
"tracing-tree",
"opentelemetry-jaeger",
"chrono",
"serde_json",
"tracing-core",
"tracing-opentelemetry",
"tracing-serde",
"tracing-subscriber",
"tracing-tree",
]
otel = [
"tracing",
"opentelemetry",
"opentelemetry-otlp",
"opentelemetry-semantic-conventions",
"opentelemetry-stdout",
"opentelemetry_sdk",
]
clap = ["dep:clap", "dep:clap_complete", "dep:clap_complete_fig"]
models = []
+2
View File
@@ -4,6 +4,8 @@
pub mod build_information;
pub mod logging;
pub mod opentelemetry;
#[cfg(feature = "clap")]
pub mod completions;
+1 -39
View File
@@ -4,17 +4,6 @@
use serde::{Deserialize, Serialize};
use std::io::IsTerminal;
#[cfg(feature = "tracing")]
pub use opentelemetry;
#[cfg(feature = "tracing")]
pub use opentelemetry_jaeger;
#[cfg(feature = "tracing")]
pub use tracing_opentelemetry;
#[cfg(feature = "tracing")]
pub use tracing_subscriber;
#[cfg(feature = "tracing")]
pub use tracing_tree;
#[derive(Debug, Default, Copy, Clone, Deserialize, PartialEq, Eq, Serialize)]
#[serde(deny_unknown_fields)]
pub struct LoggingSettings {
@@ -74,34 +63,7 @@ pub fn setup_tracing_logger() {
#[macro_export]
macro_rules! setup_tracing {
($service_name: expr) => {
use nym_bin_common::logging::tracing_subscriber::layer::SubscriberExt;
use nym_bin_common::logging::tracing_subscriber::util::SubscriberInitExt;
let registry = nym_bin_common::logging::tracing_subscriber::Registry::default()
.with(nym_bin_common::logging::tracing_subscriber::EnvFilter::from_default_env())
.with(
nym_bin_common::logging::tracing_tree::HierarchicalLayer::new(4)
.with_targets(true)
.with_bracketed_fields(true),
);
let tracer = nym_bin_common::logging::opentelemetry_jaeger::new_collector_pipeline()
.with_endpoint("http://44.199.230.10:14268/api/traces")
.with_service_name($service_name)
.with_isahc()
.with_trace_config(
nym_bin_common::logging::opentelemetry::sdk::trace::config().with_sampler(
nym_bin_common::logging::opentelemetry::sdk::trace::Sampler::TraceIdRatioBased(
0.1,
),
),
)
.install_batch(nym_bin_common::logging::opentelemetry::runtime::Tokio)
.expect("Could not init tracer");
let telemetry = nym_bin_common::logging::tracing_opentelemetry::layer().with_tracer(tracer);
registry.with(telemetry).init();
crate::opentelemetry::setup_no_otel_logger()
};
}
@@ -0,0 +1,21 @@
// Copyright 2025 - Nym Technologies SA <contact@nymtech.net>
// SPDX-License-Identifier: Apache-2.0
use opentelemetry_otlp::ExporterBuildError;
#[derive(thiserror::Error, Debug)]
pub enum TracingError {
#[error("tracing logger already initialised")]
TracingLoggerAlreadyInitialised,
// #[error("I/O error: {0}")]
// IoError(#[from] std::io::Error),
#[error("Logging error: {0}")]
TracingTryInitError(tracing_subscriber::util::TryInitError),
#[error("{0}")]
TracingExporterBuildError(#[from] ExporterBuildError),
#[error("{0}")]
TracingFilterParseError(#[from] tracing_subscriber::filter::ParseError),
}
+180
View File
@@ -0,0 +1,180 @@
pub mod error;
mod trace_id_format;
use tracing::{info, Level};
use tracing_subscriber::filter::Directive;
use tracing_subscriber::layer::SubscriberExt;
use tracing_subscriber::util::SubscriberInitExt;
use tracing_subscriber::{fmt, Layer};
use crate::logging::{default_tracing_env_filter, default_tracing_fmt_layer};
use crate::opentelemetry::error::TracingError;
use opentelemetry::trace::TracerProvider;
use opentelemetry::{global, KeyValue};
use opentelemetry_otlp::tonic_types::metadata::MetadataMap;
use opentelemetry_otlp::tonic_types::transport::ClientTlsConfig;
use opentelemetry_otlp::{WithExportConfig, WithTonicConfig};
use opentelemetry_sdk::metrics::{MeterProviderBuilder, PeriodicReader, SdkMeterProvider};
use opentelemetry_sdk::trace::{RandomIdGenerator, SdkTracerProvider};
use opentelemetry_sdk::{trace::Sampler, Resource};
use opentelemetry_semantic_conventions::resource::{DEPLOYMENT_ENVIRONMENT_NAME, SERVICE_VERSION};
use opentelemetry_semantic_conventions::SCHEMA_URL;
use tracing_opentelemetry::{MetricsLayer, OpenTelemetryLayer};
use tracing_subscriber::fmt::format::FmtSpan;
pub(crate) fn granual_filtered_env() -> Result<tracing_subscriber::filter::EnvFilter, TracingError>
{
fn directive_checked(directive: impl Into<String>) -> Result<Directive, TracingError> {
directive.into().parse().map_err(From::from)
}
let mut filter = default_tracing_env_filter();
// these crates are more granularly filtered
let filter_crates = ["defguard_wireguard_rs"];
for crate_name in filter_crates {
filter = filter.add_directive(directive_checked(format!("{crate_name}=warn"))?);
}
Ok(filter)
}
pub(crate) fn build_tracing_logger() -> Result<impl SubscriberExt, TracingError> {
let key =
std::env::var("SIGNOZ_INGESTION_KEY".to_string()).expect("SIGNOZ_INGESTION_KEY not set");
let mut metadata = MetadataMap::new();
metadata.insert(
"signoz-ingestion-key",
key.parse().expect("Could not parse signoz ingestion key"),
);
let tracer_provider = init_tracer_provider(metadata)?;
let meter_provider = init_meter_provider()?;
let tracer = tracer_provider.tracer("tracing-otel-subscriber");
let fmt_layer = fmt::layer()
.json()
.with_writer(std::io::stderr)
.with_span_events(FmtSpan::NEW | FmtSpan::CLOSE)
.with_span_list(false)
.with_current_span(true)
.event_format(trace_id_format::TraceIdFormat);
let registry = tracing_subscriber::registry()
.with(fmt_layer)
.with(granual_filtered_env()?)
.with(tracing_subscriber::filter::LevelFilter::from_level(
Level::DEBUG,
))
.with(MetricsLayer::new(meter_provider))
.with(OpenTelemetryLayer::new(tracer));
Ok(registry)
}
pub fn setup_tracing_logger() -> Result<(), TracingError> {
let stderr_layer =
default_tracing_fmt_layer(std::io::stderr).with_filter(granual_filtered_env()?);
cfg_if::cfg_if! {if #[cfg(feature = "tokio-console")] {
// instrument tokio console subscriber needs RUSTFLAGS="--cfg tokio_unstable" at build time
let console_layer = console_subscriber::spawn();
tracing_subscriber::registry()
.with(console_layer)
.with(stderr_layer)
.init();
} else {
tracing_subscriber::registry()
.with(stderr_layer)
.init();
}}
build_tracing_logger()?.init();
Ok(())
}
pub fn setup_no_otel_logger() -> Result<(), error::TracingError> {
// Only set up if not already initialized
if tracing::dispatcher::has_been_set() {
// It shouldn't be - this is really checking that it is torn down between async command executions
return Err(error::TracingError::TracingLoggerAlreadyInitialised);
}
let registry = tracing_subscriber::registry()
.with(default_tracing_fmt_layer(std::io::stderr))
.with(granual_filtered_env()?)
.with(tracing_subscriber::filter::LevelFilter::from_level(
Level::INFO,
));
registry
.try_init()
.map_err(|e| error::TracingError::TracingTryInitError(e))?;
Ok(())
}
fn resource() -> Resource {
Resource::builder()
.with_service_name("nym-node")
.with_schema_url(
[
KeyValue::new(SERVICE_VERSION, env!("CARGO_PKG_VERSION")),
KeyValue::new(DEPLOYMENT_ENVIRONMENT_NAME, "develop"),
],
SCHEMA_URL,
)
.build()
}
fn init_tracer_provider(metadata: MetadataMap) -> Result<SdkTracerProvider, TracingError> {
let endpoint = std::env::var("SIGNOZ_ENDPOINT".to_string()).expect("SIGNOZ_ENDPOINT not set");
info!("SIGNOZ_ENDPOINT = {}", endpoint);
let mut exporter_builder = opentelemetry_otlp::SpanExporter::builder()
.with_tonic()
.with_metadata(metadata)
.with_endpoint(&endpoint);
if endpoint.starts_with("https://") {
exporter_builder =
exporter_builder.with_tls_config(ClientTlsConfig::new().with_enabled_roots());
}
let exporter = exporter_builder.build()?;
let tracer = SdkTracerProvider::builder()
.with_sampler(Sampler::ParentBased(Box::new(Sampler::TraceIdRatioBased(
1.0,
))))
.with_id_generator(RandomIdGenerator::default())
.with_resource(resource())
.with_batch_exporter(exporter)
.build();
Ok(tracer)
}
fn init_meter_provider() -> Result<SdkMeterProvider, TracingError> {
let exporter = opentelemetry_otlp::MetricExporter::builder()
.with_tonic()
.with_temporality(opentelemetry_sdk::metrics::Temporality::default())
.build()?;
let reader = PeriodicReader::builder(exporter)
.with_interval(std::time::Duration::from_secs(30))
.build();
let stdout_reader =
PeriodicReader::builder(opentelemetry_stdout::MetricExporter::default()).build();
let meter_provider = MeterProviderBuilder::default()
.with_resource(resource())
.with_reader(reader)
.with_reader(stdout_reader)
.build();
global::set_meter_provider(meter_provider.clone());
Ok(meter_provider)
}
@@ -0,0 +1,88 @@
use chrono::Utc;
use opentelemetry::trace::TraceContextExt;
use opentelemetry::{SpanId, TraceId};
use serde::ser::{SerializeMap, Serializer as _};
use std::io;
use tracing::{Event, Subscriber};
use tracing_opentelemetry::OpenTelemetrySpanExt;
use tracing_serde::fields::AsMap;
use tracing_serde::AsSerde;
use tracing_subscriber::fmt::format::Writer;
use tracing_subscriber::fmt::{FmtContext, FormatEvent, FormatFields};
use tracing_subscriber::registry::LookupSpan;
pub struct WriteAdaptor<'a> {
fmt_write: &'a mut dyn std::fmt::Write,
}
impl<'a> WriteAdaptor<'a> {
pub fn new(fmt_write: &'a mut dyn std::fmt::Write) -> Self {
Self { fmt_write }
}
}
impl<'a> io::Write for WriteAdaptor<'a> {
fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
let s =
std::str::from_utf8(buf).map_err(|e| io::Error::new(io::ErrorKind::InvalidData, e))?;
self.fmt_write
.write_str(s)
.map_err(|e| io::Error::new(io::ErrorKind::Other, e))?;
Ok(s.as_bytes().len())
}
fn flush(&mut self) -> io::Result<()> {
Ok(())
}
}
pub struct TraceIdFormat;
impl<S, N> FormatEvent<S, N> for TraceIdFormat
where
S: Subscriber + for<'lookup> LookupSpan<'lookup>,
N: for<'writer> FormatFields<'writer> + 'static,
{
fn format_event(
&self,
_ctx: &FmtContext<'_, S, N>,
mut writer: Writer<'_>,
event: &Event<'_>,
) -> std::fmt::Result
where
S: Subscriber + for<'a> LookupSpan<'a>,
{
let meta = event.metadata();
let mut visit = || {
let mut serializer = serde_json::Serializer::new(WriteAdaptor::new(&mut writer));
let mut serializer = serializer.serialize_map(None)?;
serializer.serialize_entry("timestamp", &Utc::now().to_rfc3339())?;
serializer.serialize_entry("level", &meta.level().as_serde())?;
serializer.serialize_entry("fields", &event.field_map())?;
serializer.serialize_entry("target", meta.target())?;
let current_span = tracing::Span::current();
let context = current_span.context();
let span_ref = context.span();
let span_context = span_ref.span_context();
let trace_id = span_context.trace_id();
if trace_id != TraceId::INVALID {
serializer.serialize_entry("trace_id", &trace_id.to_string())?;
let span_id = span_context.span_id();
if span_id != SpanId::INVALID {
serializer.serialize_entry("span_id", &span_id.to_string())?;
}
}
serializer.end()
};
visit().map_err(|_| std::fmt::Error)?;
writeln!(writer)
}
}
-9395
View File
File diff suppressed because it is too large Load Diff