Run AWS Lambda execute async function

Allow Orchestra to run an AWS Lambda function asynchronously

Description

This job triggers an AWS Lambda function asynchronously with any given payload. To use this Integration Job, you must configure the function to be able to communicate back to Orchestra - see Self hosted Tasks. If this is not configured correctly, the Task run will be stuck in RUNNING state, and will need to be cancelled manually.

Parameters and setup

Name	Required	Data type	Restrictions
Function name	Yes	String	None
Payload	No	JSON	JSON format
Region	No	String	AWS region

Lambda configuration

The AWS Lambda function is triggered with the following event by Orchestra:

{
  "orchestra_task_run_id": "<ORCHESTRA_TASK_RUN_ID>",
  "orchestra_webhook_url": "<ORCHESTRA_WEBHOOK_URL>",
  "task_run_data": <TASK_PARAMETERS_PAYLOAD> // see above table
}

If your function is written in Python, Orchestra recommends the orchestra-sdk PyPi package to ensure the Task can communicate efficiently with our platform. Remember to install the package before running the code:

from orchestra_sdk.orchestra import OrchestraSDK

orchestra = OrchestraSDK(api_key="<ORCHESTRA_API_KEY>")

def handler(event, context):
    orchestra.configure_aws_lambda_event(event)
    # Your code here
    
# The decorator will automatically send task updates.
@orchestra.run()
def my_function(arg1, arg2=1):
    print("Running complex process")

If your function is written in another language, please refer to Self hosted Tasks documentation to see how your code can send updates back to Orchestra.

Troubleshooting

Issue	Remediation
My AWS Lambda function task is stuck in RUNNING in Orchestra?	It is likely that your AWS Lambda function has been unable to update the status of the task to Orchestra, and we have assumed it is still in RUNNING state. If this has happened, cancel the current task run, debug the issue, and then re-run the task.
My lambda function is not being started by Orchestra?	Check the AWS Lambda connection is valid and the permissions allow lambda:InvokeFunction. Other AWS policies may be denying the role access Ensure the function name and region are set correctly
I am using the orchestra-sdk package, but my lambda function is not sending updates back to Orchestra?	Make sure the API key is set correctly Use the configure_aws_lambda_event method as close to the start of the handler function as possible. If not using this method, make sure the TASK_RUN_ID and WEBHOOK_URL are set correctly You can enable debug logging in Python to see if there are any internal errors Ensure your AWS Lambda Function is running in a VPC that allows outbound internet access If you are not using the orchestra.run() decorator, make sure the update_task methods are being called correctly and the code reaches these methods as expected
My function sometimes succeeds and sometimes does not, and gets stuck in RUNNING state?	Make sure the AWS lambda function is not timing out or running out of memory - once this happens, the Task is unable to report back its status
I am expecting logs in the Orchestra platform, but they are not showing up?	Make sure send_logs=True is set in the Orchestra class when initialising the object (refer to the PyPi documentation for further instructions) Make sure your code is correctly logging to a file - for AWS Lambda functions, you can only write files to the /tmp/ directory
I am seeing logs for older task runs in my most recent task run?	AWS Lambda functions can re-use the same machine between invocations. Depending on your log setup, this means old log files still persist. Make sure to either delete log files once the function has finished, or create a new log file per invocation.