applied patch from STR-440-opentelemetry-output (#473)

STR-440 OpenTelemetry tracing exporter (#342)

* common: tried to set up otlp

This deletes Cargo.lock because we couldn't get the deps to work anyways.

* fix: bump opentelemetry stuff to 0.26

now it compiles at least-ish

* common: added more correct init for OTLP tracing output

* common, contrib: fixed some more logging stuff, added Jaeger util script

* common, stratad: reworked logging init some more to try to make it work

* contrib: fixed Jaeger init script to use a recent version

* bridge-client: fixed broken test, fixed Tokio version

* common: added `service.name` property to OTLP params

* prover-client: fixed missing `LoggerConfig`

* stratad: fixed broken tests??? idk why this is here

* tests: fixed broken logging

* btcio: fixed broken logging

---------

Co-authored-by: Trey Del Bonis <[email protected]>

Cargo.toml

@@ -37,6 +37,7 @@ members = [
   "crates/sync",
   "crates/tasks",
   "crates/test-utils",
+  "crates/tx-parser",
   "crates/util/mmr",
   "crates/util/shrex",
   "crates/vtxjmt",
@@ -46,12 +47,10 @@ members = [
   # binaries listed separately
   "bin/bridge-client",
   "bin/datatool",
-  "bin/prover-client",
   "bin/strata-cli",
   "bin/strata-client",
   "bin/strata-reth",
-  "bin/datatool",
-  "crates/tx-parser",
+  "bin/prover-client",
 
   # integration tests
   "tests",
@@ -143,6 +142,9 @@ miniscript = "12.2.0"
 mockall = "0.11"
 musig2 = { version = "0.1.0", features = ["serde"] }
 num_enum = "0.7"
+opentelemetry = "0.26"
+opentelemetry-otlp = { version = "0.26", features = ["grpc-tonic"] }
+opentelemetry_sdk = { version = "0.26", features = ["rt-tokio"] }
 parking_lot = "0.12.3"
 paste = "1.0"
 rand = "0.8.5"
@@ -208,7 +210,8 @@ threadpool = "1.8"
 tokio = { version = "1.37", features = ["full"] }
 toml = "0.5"
 tower = "0.4"
-tracing = "0.1.40"
+tracing = "0.1"
+tracing-opentelemetry = "0.27"
 tracing-subscriber = { version = "0.3.18", features = ["env-filter"] }
 uuid = { version = "1.0", features = ["v4", "serde"] }
 

bin/bridge-client/src/descriptor.rs

@@ -146,7 +146,9 @@ mod tests {
 
     #[tokio::test]
     async fn bitcoind_import_descriptors() {
-        logging::init();
+        logging::init(logging::LoggerConfig::with_base_name(
+            "bridge-client-descriptor-tests",
+        ));
         let bitcoind = BitcoinD::from_downloaded().unwrap();
         let url = bitcoind.rpc_url();
         let (user, password) = get_auth(&bitcoind);

bin/bridge-client/src/main.rs

@@ -13,12 +13,12 @@ pub(crate) mod rpc_server;
 
 use args::{Cli, OperationMode};
 use modes::{challenger, operator};
-use strata_common::logging;
+use strata_common::logging::{self, LoggerConfig};
 use tracing::info;
 
 #[tokio::main]
 async fn main() {
-    logging::init();
+    logging::init(LoggerConfig::with_base_name("strata-bridge-client"));
 
     let cli_args: Cli = argh::from_env();
 

bin/prover-client/src/main.rs

@@ -35,7 +35,9 @@ mod task;
 
 #[tokio::main]
 async fn main() {
-    logging::init();
+    logging::init(logging::LoggerConfig::with_base_name(
+        "strata-prover-client",
+    ));
     info!("Running strata prover client in dev mode");
 
     let args: Args = argh::from_env();

bin/strata-client/src/extractor.rs

@@ -218,7 +218,8 @@ mod tests {
 
     #[tokio::test]
     async fn test_extract_deposit_requests() {
-        logging::init();
+        // FIXME this is absurd, why are we doing this here?
+        logging::init(logging::LoggerConfig::with_base_name("strata-client-tests"));
 
         let (mock_db, db_config) =
             get_rocksdb_tmp_instance().expect("should be able to get tmp rocksdb instance");

bin/strata-client/src/main.rs

@@ -36,7 +36,7 @@ use strata_storage::{
 use strata_sync::{self, L2SyncContext, RpcSyncPeer};
 use strata_tasks::{ShutdownSignal, TaskExecutor, TaskManager};
 use tokio::{
-    runtime::Runtime,
+    runtime::{Handle, Runtime},
     sync::{broadcast, oneshot},
 };
 use tracing::*;
@@ -73,7 +73,15 @@ fn main() -> anyhow::Result<()> {
 }
 
 fn main_inner(args: Args) -> anyhow::Result<()> {
-    logging::init();
+    // Start runtime for async IO tasks.
+    let runtime = tokio::runtime::Builder::new_multi_thread()
+        .enable_all()
+        .thread_name("strata-rt")
+        .build()
+        .expect("init: build rt");
+
+    // Init the logging before we do anything else.
+    init_logging(runtime.handle());
 
     let config = get_config(args.clone())?;
 
@@ -100,13 +108,6 @@ fn main_inner(args: Args) -> anyhow::Result<()> {
     // initialize core databases
     let database = init_core_dbs(rbdb.clone(), ops_config);
 
-    // Start runtime for async IO tasks.
-    let runtime = tokio::runtime::Builder::new_multi_thread()
-        .enable_all()
-        .thread_name("strata-rt")
-        .build()
-        .expect("init: build rt");
-
     // Init thread pool for batch jobs.
     // TODO switch to num_cpus
     let pool = threadpool::ThreadPool::with_name("strata-pool".to_owned(), 8);
@@ -133,6 +134,8 @@ fn main_inner(args: Args) -> anyhow::Result<()> {
         genesis::init_client_state(&params, database.as_ref())?;
     }
 
+    info!("init finished, starting main tasks");
+
     let ctx = start_core_tasks(
         &executor,
         pool,
@@ -215,6 +218,29 @@ fn main_inner(args: Args) -> anyhow::Result<()> {
     Ok(())
 }
 
+/// Sets up the logging system given a handle to a runtime context to possibly
+/// start the OTLP output on.
+fn init_logging(rt: &Handle) {
+    let mut lconfig = logging::LoggerConfig::with_base_name("strata-client");
+
+    // Set the OpenTelemetry URL if set.
+    let otlp_url = logging::get_otlp_url_from_env();
+    if let Some(url) = &otlp_url {
+        lconfig.set_otlp_url(url.clone());
+    }
+
+    {
+        // Need to set the runtime context because of nonsense.
+        let _g = rt.enter();
+        logging::init(lconfig);
+    }
+
+    // Have to log this after we start the logging formally.
+    if let Some(url) = &otlp_url {
+        info!(%url, "using OpenTelemetry tracing output");
+    }
+}
+
 #[derive(Clone)]
 pub struct CoreContext {
     pub database: Arc<CommonDb>,

contrib/jaeger.sh

@@ -0,0 +1,28 @@
+#!/bin/bash
+# Script for standing up a local instance of Jaeger, so we can use it for
+# running fntests, for example.
+#
+# Adapted from: https://www.jaegertracing.io/docs/1.6/getting-started/
+
+TAG_VER=1.62.0
+
+OTLP_GRPC_PORT=4317
+WEB_PORT=16686
+
+docker run -d \
+       --name jaeger \
+       -e COLLECTOR_ZIPKIN_HTTP_PORT=9411 \
+       -e COLLECTOR_OTLP_ENABLED=true \
+       -p $OTLP_GRPC_PORT:4317 \
+       -p 4318:4318 \
+       -p 5775:5775/udp \
+       -p 6831:6831/udp \
+       -p 6832:6832/udp \
+       -p 5778:5778 \
+       -p $WEB_PORT:16686 \
+       -p 14268:14268 \
+       -p 9411:9411 \
+       jaegertracing/all-in-one:$TAG_VER
+
+echo "Set envvar: 'export STRATA_OTLP_URL=http://127.0.0.1:$OTLP_GRPC_PORT'"
+echo "Open http://localhost:$WEB_PORT/ in your browser."

crates/btcio/src/rpc/client.rs

@@ -424,7 +424,8 @@ mod test {
 
     #[tokio::test()]
     async fn client_works() {
-        logging::init();
+        logging::init(logging::LoggerConfig::with_base_name("btcio-tests"));
+
         // setting the ENV variable `BITCOIN_XPRIV_RETRIEVABLE` to retrieve the xpriv
         set_var("BITCOIN_XPRIV_RETRIEVABLE", "true");
         let bitcoind = BitcoinD::from_downloaded().unwrap();

crates/common/Cargo.toml

@@ -4,5 +4,9 @@ name = "strata-common"
 version = "0.1.0"
 
 [dependencies]
+opentelemetry.workspace = true
+opentelemetry-otlp.workspace = true
+opentelemetry_sdk.workspace = true
 tracing.workspace = true
+tracing-opentelemetry.workspace = true
 tracing-subscriber.workspace = true