« Running Locally

Running on AWS Step Functions »

Running on AWS

Build and Deploy Lambda Function Code

Build .ZIP

Build a .zip file containing you application code. For example, to build the aws-lambda-fsm example applications you can do the following:

$ workon aws-lambda-fsm-workflows
(aws-lambda-fsm)[aws-lambda-fsm-workflows (master)]$ make build
  ...
  adding: yaml/serializer.py (deflated 78%)
  adding: yaml/tokens.py (deflated 78%)

If you have written a custom application that brings in aws-lambda-fsm as a module, you will have to ensure that pyyaml and (optionally) python-memcached are included in your .zip file.

Upload to S3

Configure AWS Services

Setup ~/.aws/credentials

$ cat ~/.aws/credentials
# Will expire on 2016-03-02T17:22:44Z
aws_access_key_id = ...
aws_secret_access_key = ...
aws_security_token = ...
aws_session_token = ...

Setup ~/.aws/config

$ cat ~/.aws/config
[default]
region=us-east-1
output=json

Create Kinesis Stream

$ workon aws-lambda-fsm
$ python tools/create_kinesis_stream.py --kinesis_num_shards=##

Create DynamoDB Tables

$ workon aws-lambda-fsm
$ python tools/create_dynamodb_table.py --dynamodb_read_capacity_units=## --dynamodb_write_capacity_units=##

Create Memcache/Redis

TODO:

Create and Configure the Lambda Function

Code

Handler

Configure the Lambda function handler as main.lambda_handler

Event Sources

Configure the Lambda function with three event sources

1. Kinesis Stream

1. DynamoDB Update

1. CloudWatch Timer

IAM

Configure the Lambda function with a role like the following:

Start a State Machine

You can use the start_state_machine.py script to start machines on AWS as well:

$ workon aws-lambda-fsm-workflows
$ python tools/start_state_machine.py --machine_name=tracer

« Running Locally

Running on AWS Step Functions »