diff --git a/.changeset/forward-worker-instance-name.md b/.changeset/forward-worker-instance-name.md
new file mode 100644
index 00000000000..368645255e4
--- /dev/null
+++ b/.changeset/forward-worker-instance-name.md
@@ -0,0 +1,14 @@
+---
+"trigger.dev": patch
+---
+
+Forward `TRIGGER_WORKER_INSTANCE_NAME` into the managed run process env
+
+The Kubernetes workload manager already injects `TRIGGER_WORKER_INSTANCE_NAME`
+(= `spec.nodeName`) into the run-controller container via the downward API, but
+the managed run worker is forked with an explicit env allowlist
+(`RunnerEnv.gatherProcessEnv()`) that dropped it. As a result the run process
+could not attach the node as an OpenTelemetry `host.name` resource attribute, so
+spans/logs exported to an off-node OTLP collector had no host (Datadog tags these
+`issue_type:empty_hostname`). This forwards the value so `trigger.config.ts`
+telemetry (or the worker itself) can surface the node/host.
diff --git a/packages/cli-v3/src/entryPoints/managed/env.ts b/packages/cli-v3/src/entryPoints/managed/env.ts
index 4da8361e863..18af293cf22 100644
--- a/packages/cli-v3/src/entryPoints/managed/env.ts
+++ b/packages/cli-v3/src/entryPoints/managed/env.ts
@@ -236,6 +236,11 @@ export class RunnerEnv {
       OTEL_EXPORTER_OTLP_ENDPOINT: this.OTEL_EXPORTER_OTLP_ENDPOINT,
       TRIGGER_OTEL_EXPORTER_OTLP_ENDPOINT: this.OTEL_EXPORTER_OTLP_ENDPOINT,
       UV_USE_IO_URING: this.UV_USE_IO_URING,
+      // Forward the worker instance name (spec.nodeName on k8s) into the run
+      // process so it can be attached as an OTEL host/node resource attribute.
+      // Without this, run spans/logs exported to an off-node OTLP collector
+      // have no host and Datadog tags them issue_type:empty_hostname.
+      TRIGGER_WORKER_INSTANCE_NAME: this.TRIGGER_WORKER_INSTANCE_NAME,
     };
 
     // Filter out undefined values