# Socket error after 60s

**URL:** <https://forum.langchain.com/t/socket-error-after-60s/448>\
**Category:** Agent Server\
**Tags:** cloud\
**Created:** [July 14, 2025, 6:02pm UTC](https://forum.langchain.com/t/socket-error-after-60s/448 "2025-07-14T18:02:14Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![lukanarrative](https://avatars.discourse-cdn.com/v4/letter/l/f04885/32.png) [@lukanarrative](https://forum.langchain.com/u/lukanarrative)\
**Post date:** [July 14, 2025, 6:02pm UTC](https://forum.langchain.com/t/socket-error-after-60s/448/1 "2025-07-14T18:02:14Z")

</div>

Hi,

We are encountering an issue when invoking agents on the langgraph platform.

The agents finish successfully - we can see that in the Langsmith logs, but the sdk requests we make fail after 60s:

We create a run with this code in typescript:

```auto
try {
      // Create and start a new run
      const run = await this.client.runs.create(threadId, assistantID, {
        input,
      });

      // Wait for the run to complete and gather the response
      const response = await this.client.runs.join(threadId, run.run_id);

      this.logger.log(`Received response from LangChain Graph: ${JSON.stringify(response)}`);

      // Return the response as a JSON object
      return response as T;
    } catch (error) {
      this.logger.error(`Error during LangChain Graph request: ${JSON.stringify(error)}`);
      throw error;
    }

```

The run is successfully created, we can see that in langgraph platform and on Langsmith.

After 60s, we get this error (I’ve just changed my ip address with ‘\*’).

```auto
{
  "cause": {
    "name": "SocketError",
    "code": "UND_ERR_SOCKET",
    "socket": {
      "localAddress": " **.**. **.**",
      "localPort": 49588,
      "remoteAddress": "35.230.68.103",
      "remotePort": 443,
      "remoteFamily": "IPv4",
      "bytesWritten": 1346,
      "bytesRead": 430
    }
  }
}

```

The run successfully finishes on the langgraph platform - we can see the output in langsmith logs.

Any ideas why this is happening?

This is happening in environemnt with deployment id `36ce7f06-2950-4fb9-8295-da9a24c7a789`

---

<div class="post-metadata">

**Author:** ![hari](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/hari/32/187_2.png) [@hari](https://forum.langchain.com/u/hari)\
**Post date:** [July 15, 2025, 5:26am UTC](https://forum.langchain.com/t/socket-error-after-60s/448/2 "2025-07-15T05:26:37Z")

</div>

Hi @lukanarrative, can you try setting a longer timeout on your client instance? Requests to /threads/{thread\_id}/runs/{run\_id}/join may take longer than 60 seconds

---

<div class="post-metadata">

**Author:** ![lukanarrative](https://avatars.discourse-cdn.com/v4/letter/l/f04885/32.png) [@lukanarrative](https://forum.langchain.com/u/lukanarrative)\
**Post date:** [July 15, 2025, 12:08pm UTC](https://forum.langchain.com/t/socket-error-after-60s/448/3 "2025-07-15T12:08:25Z")

</div>

Found the core issue → our service’s VPN was silently killing requests longer than 60s, that caused this issue 👍  
Nothing wrong on the LangGraph’s side!
