Motivation

Lambda records a CloudWatch (CW) error metric upon every failed Lambda invocation but LocalStack does not record the error metric anymore since the Lambda invocation loop rework. @steffyP recently added a test for this behavior in #9653 discovered in a Pro sample.

Errors – The number of invocations that result in a function error. Function errors include exceptions that your code throws and exceptions that the Lambda runtime throws. The runtime returns errors for issues such as timeouts and configuration errors. To calculate the error rate, divide the value of Errors by the value of Invocations. Note that the timestamp on an error metric reflects when the function was invoked, not when the error occurred.

see AWS CW metrics

Changes

• Fix the regression and submit a CloudWatch metric upon failed Lambda invocation
• Update some Lambda docs (nit)

Discussion

• Exceptions that are not caught in localstack.services.lambda_.invocation.version_manager.LambdaVersionManager.invoke (i.e., all exceptions except StatusErrorException) won't be recorded as errors. For example, exceptions from the Lambda control plane (e.g., TooManyRequestsException) should be recorded in other metrics (e.g., Throttles).

Background

looks like a bug introduced in the Lambda invocation loop rework: https://github.com/localstack/localstack/pull/8970/files#diff-056f8f1dc9188bf87729941f6be9805c0caf16189dafa882f16fb259676f79daL727

similar to the record_invocation case here:

https://github.com/localstack/localstack/pull/8970/files#diff-056f8f1dc9188bf87729941f6be9805c0caf16189dafa882f16fb259676f79daR220

This should be fixed in the version_manager.py::invoke