(fix) Prevent Datadog timeout errors in logs

.changesets/fix_bryn_datadog_timeout.md

@@ -0,0 +1,12 @@
+### Prevent Datadog timeout errors in logs ([Issue #2058](https://github.com/apollographql/router/issue/2058))
+
+The telemetry Datadog exporter uses connection pooling by default. However, these pools connections are frequently closed leading to errors in the router logs.
+
+For example:
+```
+2024-07-19T15:28:22.970360Z ERROR  OpenTelemetry trace error occurred: error sending request for url (http://127.0.0.1:8126/v0.5/traces): connection error: Connection reset by peer (os error 54)
+```
+
+The pool timeout for the Datadog exporter is now set to 1 millisecond, which will greatly reduce the frequency that this occurs.
+
+By [@BrynCooke](https://github.com/BrynCooke) in https://github.com/apollographql/router/pull/5692

apollo-router/src/plugins/telemetry/tracing/datadog.rs

@@ -2,6 +2,7 @@
 
 use std::fmt::Debug;
 use std::fmt::Formatter;
+use std::time::Duration;
 
 use ahash::HashMap;
 use ahash::HashMapExt;
@@ -196,7 +197,13 @@ impl TracingConfigurator for Config {
                     .expect("cargo version is set as a resource default;qed")
                     .to_string(),
             )
-            .with_http_client(reqwest::Client::builder().build()?)
+            .with_http_client(
+                reqwest::Client::builder()
+                    // https://github.com/open-telemetry/opentelemetry-rust-contrib/issues/7
+                    // Set the idle timeout to something low to prevent termination of connections.
+                    .pool_idle_timeout(Duration::from_millis(1))
+                    .build()?,
+            )
             .with_trace_config(common)
             .build_exporter()?;