AWS Compute Blog
Testing AWS Lambda functions with AWS SAM remote invoke
Developers are taking advantage of event driven architecture (EDA) to build large distributed applications. To build these applications, developers are using managed service like AWS Lambda, AWS Step Functions, and Amazon EventBridge to handle compute, orchestration, and choreography. Since these services run in the cloud, developers are also looking for ways to test in the cloud. With this in mind, AWS SAM is adding a new feature to the AWS SAM CLI called remote invoke.
AWS SAM remote invoke enables developers to invoke a Lambda function in the AWS Cloud from their development environment. The feature has several options for identifying the Lambda function to invoke, the payload event, and the output type.
Using remote invoke
To test the remote invoke feature, there is a small AWS SAM application that comprises two AWS Lambda functions. The TranslateFunction takes a text string and translates it to the target language using the AI/ML service Amazon Translate. The StreamFunction generates data in a streaming format. To run these demonstrations, be sure to install the latest AWS SAM CLI.
To deploy the application, follow these steps:
- Clone the repository:
git clone https://github.com/aws-samples/aws-sam-remote-invoke-example
- Change to the root directory of the repository:
cd aws-sam-remote-invoke-example
- Build the AWS Lambda artifacts (use the –use-container option to ensure Python 3.10 and Node 18 are present. If these are both set up on your machine, you can ignore this flag):
sam build --use-container
- Deploy the application to your AWS account:
sam deploy --guided
- Name the application “remote-test” and choose all defaults.
AWS SAM can now remotely invoke the Lambda functions deployed with this application. Use the following command to test the TranslateFunction:
sam remote invoke --stack-name remote-test --event '{"message":"I am testing the power of remote invocation", "target-language":"es"}' TranslateFunction
This is a quick way to test a small event. However, developers often deal with large complex payloads. The AWS SAM remote invoke function also allows an event to be passed as a file. Use the following command to test:
sam remote invoke --stack-name remote-test --event-file './events/translate-event.json' TranslateFunction
With either of these methods, AWS SAM returns the response from the Lambda function as if it were called from a service like Amazon API Gateway. However, AWS SAM also offers the ability to get the raw response as returned from the Python software development kit (SDK), boto3. This format provides additional information such as the version that you invoked, if any retries were attempted, and more. To retrieve this output, run the invocation with the additional –output parameter with the value of json.
sam remote invoke --stack-name remote-test --event '{"message": "I am testing the power of remote invocation", "target-language": "es"}' --output json TranslateFunction
It is also possible to invoke Lambda functions that are not created in AWS SAM. Using the name of a Lambda function, AWS SAM can remotely invoke any Lambda function that you have permission to invoke. When you deployed the sample application, AWS SAM prints the name of the Lambda function in the console. Use the following command to print the output again:
sam list stack-outputs --stack-name remote-test
Using the output for the TranslateFunctionName, run:
sam remote invoke --event '{"message": "Testing direct access of the function", "target-language": "fr"}' <TranslateFunctionName>
Lambda recently added support from streaming responses from Lambda functions. Streaming functions do not wait until the entire response is available before they respond to the client. To show this, the StreamFunction generates multiple chunks of text and sends them over a period of time.
To invoke the function, run:
sam remote invoke --stack-name remote-test StreamFunction
Extending remote invoke
The AWS SDKs offer different options when invoking Lambda functions via the Lambda service. Behind the scenes, AWS SAM is using boto3 to power the remote invoke functionality. To make full use of the SDK options for Lambda function invocation, the AWS SAM offers a —parameter flag that can be used multiple times.
For example, you may want to run an invocation as a dry run only. This type of invocation tests Lambda’s ability to invoke the function based on factors like variable values and proper permissions. The command looks like the following:
sam remote invoke --stack-name remote-test --event '{"message": "I am testing the power of remote invocation", "target-language": "es"}' --parameter InvocationType=DryRun --output json TranslateFunction
In a second example, I want to invoke a specific version of the Lambda function:
sam remote invoke --stack-name remote-test --event '{"message": "I am testing the power of remote invocation", "target-language": "es"}' --parameter Qualifier='$LATEST' TranslateFunction
If you need both options:
sam remote invoke --stack-name remote-test --event '{"message": "I am testing the power of remote invocation", "target-language": "es"}' --parameter InvocationType=DryRun --parameter Qualifier='$LATEST' --output json TranslateFunction
Logging
When developing distributed applications, logging is a critical tool to trace the state of a request across decoupled microservices. AWS SAM offers the sam logs functionality to help view aggregated logs and traces from Amazon CloudWatch and AWS X-Ray, respectively. However, when testing individual functions, developers want contextual logs pinpointed to a specific invocation. The new remote invoke function provides these logs by default. Returning to the TranslateFunction, run the following command again:
sam remote invoke --stack-name remote-test --event '{"message": "I am testing the power of remote invocation", "target-language": "es"}' TranslateFunction
The remote invocation returns the response from the Lambda function, any logging from within the Lambda function, followed by the final report from the Lambda service about the invocation itself.
Combining remote invoke with AWS SAM Accelerate
Developers are constantly striving to remove complexity and friction and improve speed and agility in the development pipeline. To help serverless developers towards this goal, the AWS SAM team released a feature called AWS SAM Accelerate. AWS SAM Accelerate is a series of features that move debugging and testing from the local machine to the cloud.
To show how AWS SAM Accelerate and remote invoke can work together, follow these steps:
- In a separate terminal, start the AWS SAM sync process with the watch option:
sam sync --stack-name remote-test --use-container --watch
- In a second window or tab, run the remote invoke function:
sam remote invoke --stack-name remote-test --event-file './events/translate-event.json' TranslateFunction
The combination of these two options provides a robust auto-deployment and testing environment. During iterations of code in the Lambda function, each time you save the file, AWS SAM syncs the code and any dependencies to the cloud. As needed, the remote invoke is then run to verify the code works as expected, with logging provided for each execution.
Conclusion
Serverless developers are looking for the most efficient way to test their applications in the AWS Cloud. They want to invoke an AWS Lambda function quickly without having to mock security, external services, or other environment variables. This blog shows how to use the new AWS SAM remote invoke feature to do just that.
This post shows how to invoke the Lambda function, change the payload type and location, and change the output format. It explains using this feature in conjunction with the AWS SAM Accelerate features to streamline the serverless development and testing process.
For more serverless learning resources, visit Serverless Land.