When pressing the Execute workflow button the connection is lost, and there’s no error in the server logs.
But if a request comes in via the webhook, it runs successfully.
Hi @Punosie this is a known issue many people faced that connection lost, the fix for this is to set a custom health endpoint as GCP reserve uses /healthz so that should be like /health
N8N_ENDPOINT_HEALTH=health
Let me know if that works.
This is the response received on the endpoint after update.
But still cannot run the workflows manually, as soon as the button is clicked still the same error:
Please share the Docker Compose variables or the config itself. But remove any sensitive info from them.
sounds like the manual execute is using a different execution process than the webhook. on GCP Cloud Run the connection sometimes drops if the process takes too long or if there’s a network timeout. check your n8n logs – look for connection reset or timeout errors around the execution attempt. also verify that your Cloud Run instance has enough memory and that the execution timeout isn’t too strict
sounds like the manual execute is using a different execution process than the webhook. on GCP Cloud Run the connection sometimes drops if the process takes too long or if there’s a network timeout. check your n8n logs – look for connection reset or timeout errors around the execution attempt. also verify that your Cloud Run instance has enough memory and that the execution timeout isn’t too strict
