Child pages
  • API Error Retries
Skip to end of metadata
Go to start of metadata

Error Response Handling

This section details a simple paradigm for error response handling.

Applications based on request/response pairs may face a myriad of issues because of network and internet connectivity causing request timeouts due to network throttling and overloading. Many hardware problems, such as load-balancer faults and switch timeouts, can also cause transparent request timeouts. For these reasons, an application using the Apstrata database service should implement a request retry strategy.

The following is an example of a simple strategy for exponential back-off in request retries. The back-off strategy is dependent on the application’s business needs. It stops retrying a request after 5 tries in order to avoid network congestion:

nRetries = 5
for (i=0; i<nRetries; i++) {
	response = Make the request
	if ((response is success) OR (response is a client error)) {
	return the response
// Response was a server error
// Delay the next request by waiting
wait(4 ^ i * [Random number between 1 and 2] * 100 milliseconds)

Client/Server Errors

Do not retry client errors. Client errors indicate that Apstrata found a problem with the client request and the application should address the issue before submitting the request again. You should, however, retry requests if server errors occur.

  • No labels