This sounds like it's simply the HTTP request deadline of the actual client request, and not your application code returning deadline exception errors. The client requests have a default deadline (specific to the library performing the HTTP request) that it will wait for your application to return a response. If your application is too busy/slow the client will give up waiting and time out.
There are a number of things which could be taking up response time (e.g network latency) so it is recommended to ensure your APIs always respond fast. This really comes down to your scaling settings in your
app.yaml or
appengine-web.xml, and your actual code. If requests are timing out that means either the requests are waiting too long for available CPU time on an available instance, or your code takes too long to finish and return a response.
It is therefore recommended to use the default Automatic scaling settings as these are fine-tuned to provide proper request handling depending on your traffic patterns. You should also avoid slow code initialization times (e.g loading dynamic libraries), and offload all long-running non-client facing work to a different
backend service as this will ensure your front-end client-facing instances are quick to start and always have available CPU to handle requests.
Of course you can always increase the client-side request deadline, but this is not always an available option (e.g if the client requests are not in your control).
- Note that Google Groups is for general product discussions and not for technical support. If you require further technical support with serving your requests it is recommended to post
detailed questions to
Stack Overflow using a supported Cloud tag.