-
Notifications
You must be signed in to change notification settings - Fork 297
Gather_Diagnostic_Info
When contacting AWS Support or raising an issue/pull-request on this repository, you may be asked to provide diagnostic information to aid in troubleshooting your error or validating your proposed fix. You can use the below steps to gather some of the most common information.
- Athena Query ID that exhibited the error.
- AWS Region you ran the query in.
- Name of the affected connector(s)/UDF(s) and their respective versions. (versions can be found in Serverless Application Repository or by the release tag you built from this repository)
- Copy of any error messages or stack traces.
- Description of the problem.
- Description of the expected behavior.
- Cloudwatch Logs for the relevant Lambda invocations (details can be found below)
- (required only for custom connectors) Output of Connector Validator
It is often helpful to restrict the level of parallelism Athena can achieve when calling your connector. By restricting your Lambda function to use a single (1) invocation reserved capacity you can force all relevant longs to go into a single log stream in Cloudwatch logs. This can make it much easier to see the call flow and spot errors / exceptions in your connector. This is especially useful while building/testing new connectors in conjunction with the Connector Validator. Keep in mind that when you set the reserved concurrency in Lambda you will affect all callers of that Lambda function, not just yourself.
Alternatively, you can use Athena to query Cloudwatch Logs and search/export all relevant logs.