When your function results in 4xx response codes, Lambda registers the invocation as an error and may retry. There are many reasons this might happen, including:
Possible Solutions
- Unhandled exceptions causing the code to exit
- Out-of-memory exceptions
- Timeouts
- Permissions errors
These could be caused by errors in your code. Logs can be extremely helpful in diagnosing these issues. Lambda stores the logs emitted by your function to Cloudwatch Logs. Dead letter queues can also be helpful because they allow you to look at failed event payloads.
Dead Letter Errors
When events are sent asynchronously to Lambda, any event that fails twice will be sent to a configured dead letter queue; however, if there are any errors in sending the event payload to the dead letter queue, Cloudwatch registers a dead letter error. Blue Matador monitors this metric so you don’t miss any errors. If your function is producing dead letter errors, the following may be possible:
Possible Solutions
- Permissions errors
- Throttles from downstream services
- Misconfigured resources
- Timeouts
Function Error Rate
By monitoring the error rate of AWS Lambda functions, we provide crucial insights into potential issues within your serverless architecture. A spike in error rates could signify various underlying issues requiring attention. These may include misconfigured Lambda function permissions, which could restrict necessary access and lead to failures.
Throttling from downstream services could also overwhelm the Lambda function, resulting in errors under heavy load conditions. Misconfigured resources, such as insufficient memory allocation or improper configuration settings, might also contribute to increased error rates. Here's a list of possible causes for increased AWS Lambda function error rates:
Possible Solutions
- Increase memory allocation or adjust concurrency settings to ensure Lambda functions have adequate resources to handle incoming requests efficiently.
- Review and update event source configurations to ensure they provide valid input data and are properly integrated with Lambda functions.
- Bugs, logic errors, or unhandled exceptions within the Lambda function code can lead to unexpected failures and elevated error rates.
