easier, simpler, faster, better
Seven months ago I published the lambdash AWS Lambda Shell Hack that lets you run shell commands to explore the environment in which AWS Lambda functions are executed.
I also posted samples of command output that show fascinating properties of the AWS Lambda runtime environment.
In the last seven months, Amazon has released new features and enhancements that have made a completely new version of lambdash possible, with many benefits including:
Ability to use AWS CloudFormation to create all needed resources including the AWS Lamba function and the IAM role.
Ability to create AWS Lambda functions by referencing a ZIP file in an S3 bucket.
Simpler IAM role structure.
Increased AWS Lamba function memory limit, with corresponding faster execution.
Ability to invoke an AWS Lambda function synchronously.
This last point means that we no longer need to put the shell command output into an S3 bucket and poll the bucket from the local host. Instead, we can simply return the shell command output directly to the client that invoked the AWS Lambda function.
The above have made the lambdash code much simpler, much easier to intstall, and much, much faster to execute and get results.
You can browse the source here:
There are three easy steps to get
1. CloudFormation Stack
Option 1: Here are sample steps to create the lambdash AWS Lambda function and to use a local command to invoke the function and output the results of commands run inside of Lambda
git clone email@example.com:alestic/lambdash.git cd lambdash ./lambdash-install
lambdash-install script runs the aws-cli command
cloudformation create-stack passing in the template file to create
the AWS Lambda function in a CloudFormation stack.
The above assumes that you have installed aws-cli and have appropriate credentials configured.
Option 2: You may use the AWS Console to create a lambdash CloudFormation stack by pressing this button:
Accept all the defaults, confirm the IAM role creation (after reading the CloudFormation template and verifying that I am not doing anything malicious), and perhaps add a Tag to help identify the lambdash CloudFormation stack.
2. Environment Variable
Since the CloudFormation stack creates the AWS Lambda function with a
unique name, you need to find out what this name is before you can
invoke it with the
If you ran the
lambdash-install command, it printed the export
statement you should use.
If you used the AWS Console, click on the lambdash CloudFormation stack’s [Output] tab and copy the export command listed there.
It will look something like this, with your own unique 12-character suffix:
Run this in your current shell and, perhaps, add it to your
$HOME/.bashrc or equivalent.
3. Local lambdash Program
The previous step installs the AWS Lambda function in the AWS environment. You also need a complementary local command that will invoke the function with your requested command line then receive and print the stdout and stderr content.
This is the
lambdash program, which is now a small Python script
that uses boto3.
You can either use the
lambdash program in the GitHub repo you
cloned above, or download it directly:
sudo curl -so/usr/local/bin/lambdash \ https://raw.githubusercontent.com/alestic/lambdash/master/lambdash sudo chmod +x /usr/local/bin/lambdash
This Python program requires boto3, so install it using your favorite method. This worked for me:
sudo -H pip install boto3
Now you’re ready to run shell commands on AWS Lambda.
You can now execute shell commands in the AWS Lambda environment and see the output. This command shows us that Amazon has upgraded the AWS Lambda environment from Amazon Linux 2014.03 when it was launched, to 2015.03 today:
$ lambdash cat /etc/issue Amazon Linux AMI release 2015.03 Kernel \r on an \m
Nodejs has been upgraded from v0.10.32 to v0.10.36
$ lambdash node -v v0.10.36
Here’s a command I use to occasionally check in on changes in the Amazon’s awslambda nodejs framework that runs our Lambda functions:
mkdir awslambda-source lambdash tar cvzf - -C /var/runtime/node_modules/awslambda . | tar xzf - -C awslambda-source
For example, the most recent change was to “log only 256K of errorMessage into customer’s cloudwatch”. Good to know.
Deleting the lambdash CloudFormation stack removes all resources including the AWS Lambda function and the IAM role. You can do this by running this command in the GitHub repo:
Or, you can delete the lambdash CloudFormation stack in the AWS Console.