Amazon Marketplace Web Service (Amazon MWS) deals with errors using the standard HTTP
Response status codes. A message is normally sent with the error code to explain
more about the error so that the problem it can be fixed or reported.
These error codes are common to all Amazon MWS API
sections. For error codes that are specific to an API section, see that API's
corresponding error codes section reference.
Table 1. Common HTTP error status codes
|HTTP status code
|There was an error reading the input stream.
|An invalid parameter value was used, or the request size
exceeded the maximum accepted size, or the request
|Access was denied.
|An invalid AWSAccessKeyId value was
|The signature used does not match the server's calculated
|An invalid API section or operation value was used, or an
invalid path was used.
|There was an internal service failure.
|The total number of requests in an hour was exceeded.
|The frequency of requests was greater than allowed.
When you get an error
Each error is accompanied with enough information to help you research a fix for the error on
the client side. When your research can't resolve the problem, you can request
assistance from Amazon MWS technical support.
However, before calling Amazon MWS technical
support, please do the following:
- Review the request that preceded the problem. Then, look for the problem
described in the error message. Resolve it, if possible.
- Record the RequestId and Timestamp
of the best example request that generated the error.
- If you cannot resolve the problem and need to call for technical support,
Amazon needs the RequestId and
Timestamp to locate the specific instance of your
issue in our logs.
Important: You should record and retain for 30 days the
RequestId and Timestamp for every
request that you submit to Amazon MWS. Doing this
enables Amazon to diagnose and fix any problems that you might encounter,
helping to improve your experience using Amazon MWS. Without this information, requests to Amazon MWS support will take longer to resolve.
The common response to a 500 or 503 service error is to try the request again. Such service
errors are usually only temporary and will resolve themselves. If you want to
retry an operation call after receiving a 500 or 503 error, you can immediately
retry after the first error response. However, if you want to retry multiple
times, Amazon recommends that you implement an "exponential backoff"
approach (i.e. pausing between retrys), with up to four retries. Then, log the
error and proceed with a manual follow-up and investigation. For example, you
can time your retries with the following time spacing: 1s, 4s, 10s, 30s. The
actual backoff times and limit will depend upon your business processes.