AWS Serverless Error Handling Summary by Execution Model
Error Handling Summary by Execution Model
The table bellow compares error handling by execution model.
| API Gateway | Amazon SNS | Kinesis Data Streams | SQS Queue | |
|---|---|---|---|---|
| Timeout considerations | API Gateway has a 30-second timeout | Requests are sent to Lambda, where they are queued and executed | Configured by the retention period of the records | Configured by the visiy timeout |
| Retries | There are no built-in retries | Has built-in retries | Configured by "Maximum Retry Attempts" number | "MaxReveveCount" is used tit the retry number |
| Error handling | You must implement error handling | Use Lambda on-failure destination or a dead-letter queue | Configure Lambda on-failure destination with the "Maximum Retry Attempts" number | Use a dead-letter queue to handle failed messages |
Related reads:
Error Handling and Automatic Retries in AWS LambdaConfiguring Error Handling for Asynchronous Invocation
Handle Lambda Errors in API Gateway
Amazon API Gatewaits and Important Notes
Amazon SNS Message ery Retries
Using AWS Lambda with Amazon Kinesis: Error Handling
Kinesis Data Streams: Changing the Data Retention Period
Amazon SQS Visiy Timeout
