We removed our free Sandbox April 25th.
You can read more on our blog.
So...You’re seeing the dotCloud active health check looking to make sure that your service is up. There is no way for you to disable it, but you can prevent it and you can handle it!
You can prevent this probe by making sure your app doesn’t return any 5xx errors unless there really is a problem. As soon as our router software, hipache sees a 5xx error, it will ask the active health check to double-check the health of the application. That causes us to do the HEAD on /CloudHealthCheck. If we get anything back within 3 seconds AND it is not a 5xx error, then we consider the service to be up and we keep it in the load-balanced rotation. If there is no response or we get a 5xx then we take that service out of rotation. In either case, we start checking /CloudHealthCheck once every 3 seconds to see if there is a problem. Since there are multiple dotCloud gateways each doing the check, you can see this check more than once per second. We’ll stop the probes after 30 minutes of no 5xx’s and no delayed responses.
There are several ways to handle this probe:
- Do nothing. A 404 is not a problem and will not remove your service from the load-balanced rotation.
- Modify your nginx to immediately return a 200.
- Add a small file or handler to your app at /CloudHealthCheck
- Actually do more tests, like check your database connectivity, and return an appropriate response within 3 seconds. Remember, we only start checking /CloudHealthCheck if we already saw a 5xx error, so maybe there really is a problem