[Question]: Code component read timeout

[SpongeKim666]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

Describe your problem

Every time I run a code component, a timeout error occurs, even with the default component. I've checked Docker configurations: services start normally (logs confirm), containers are Up and healthy, ports are mapped correctly (e.g., 0.0.0.0:9385->9385/tcp). DNS resolution (nslookup, ping) and port connectivity (telnet) work, yet timeouts persist. @Duso

Error coede:
Error: Internal error in sanbox: HTTPConnectionPool(host='sandbox-executor-manager', port=9385): Read timed out. (read timeout=10).

[dosubot]

Your timeout error—HTTPConnectionPool(host='sandbox-executor-manager', port=9385): Read timed out. (read timeout=10)—is likely caused by the sandbox-executor-manager service not responding within the default 10-second timeout window. This timeout is configurable and managed by the SANDBOX_TIMEOUT environment variable. If your code component or the container startup takes longer than 10 seconds, the request will fail with a timeout, even if the Docker containers appear healthy and ports are mapped correctly.

To address this:

1. Check your environment variables (especially in your Docker setup) and try increasing SANDBOX_TIMEOUT to a higher value (for example, 30s or 60s). The service parses values like 30s, 1m, or 1m30s—invalid or missing values default to 10 seconds.
   See how the timeout is parsed and used here and here.

2. After updating the timeout, restart your containers to ensure the new value is picked up.

3. If the issue persists, check the logs of the sandbox-executor-manager container for errors or warnings. Even if the container is "healthy," the service inside may be failing due to internal errors, resource limits, or misconfiguration. The status of a Docker container does not always reflect the health of the service running inside it—network failures, incorrect port numbers, or DNS issues can still cause problems even when containers are up and healthy (source).

4. If you see repeated failures, monitor the logs for clues about slow container allocation or code execution. The codebase uses a 10-second timeout for both container creation and code execution by default, so slow startup or heavy workloads can trigger timeouts (source).

If you have already increased the timeout and confirmed network connectivity, but the problem remains, there may be an issue with the code component itself or with resource constraints on your host. In that case, reviewing the logs for specific error messages during the timeout window will help pinpoint the root cause.

To reply, just mention my name @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[yongtenglei]

Did you execute the following commands?

docker pull infiniflow/sandbox-base-nodejs:latest
docker pull infiniflow/sandbox-base-python:latest

May I see the output of docker ps, docker logs -f sandbox-executor-manager? Or docker logs -f ragflow-sandbox-executor-manager in case there is no container naed sandbox-executor-manager, due to a recent change that changed the container name.

[SpongeKim666]

Did you execute the following commands?

docker pull infiniflow/sandbox-base-nodejs:latest
docker pull infiniflow/sandbox-base-python:latest
May I see the output of docker ps, docker logs -f sandbox-executor-manager? Or docker logs -f ragflow-sandbox-executor-manager in case there is no container naed sandbox-executor-manager, due to a recent change that changed the container name.

I have executed the following commands, but still don‘t work. This is the log file of yesterday (black) (before I pull the two images) and today (white) (after I pull the two images). @yongtenglei

[yongtenglei]

I see... You didn't install the gVisor, see README and sandbox.