• Configuration
  • SSM parameters
  • Environment variables
• Testing
• Deployment

A Go based Amazon Lambda function to periodically delete stale tweets, complete with:

• Deployment configuration via Serverless Application Model (SAM)
• Secure Twitter API credential lookup via Simple Systems Manager (SSM) Parameter Store
• Local Lambda execution environment testing via aws-sam-cli

Heavily inspired by ephemeral, by @vickylai. All of the tweet deleting functionality was redone from scratch as an exercise to get familiar with Go and the AWS SDK for Go. Also, it was a good excuse to take a closer look at SAM.

All of the application configuration is provided through a mixture SSM parameters and environment variables. The sensitive ones come through the former, and the rest through the latter.

• /neuralyzer/twitter/accessToken: Twitter API access token
• /neuralyzer/twitter/accessTokenSecret: Twitter API access token secret
• /neuralyzer/twitter/consumerKey: Twitter API consumer key
• /neuralyzer/twitter/consumerKeySecret: Twitter API consumer key secret

• NEURALYZER_TWEET_AGE_THRESHOLD: Tweet age threshold for deleting tweets (default: 2190h)

Once the SSM parameters above exist, install sam and its prerequisites (Docker, Python 2.7).

Note: Be careful not to accidentally delete all of your tweets by invoking the function locally!

Next, use the testacc target of the Makefile to build a Linux compatible binary and execute it within a container image that mimics the Amazon Lambda execution environment for Go.

$ make testacc -e AWS_PROFILE=personal
GOOS=linux go build -o main main.go
sam local generate-event schedule | sam local invoke NeuralyzerFunction
2018-05-29 23:05:59 Reading invoke payload from stdin (you can also pass it from file with --event)
2018-05-29 23:06:00 Invoking main (go1.x)
2018-05-29 23:06:00 Found credentials in shared credentials file: ~/.aws/credentials

Fetching lambci/lambda:go1.x Docker container image......
2018-05-29 23:06:00 Mounting /Users/hector/.go/src/github.com/hectcastro/neuralyzer as /var/task:ro inside runtime container
START RequestId: 747fc7fe-5471-13ca-a7ae-8cbeb0335031 Version: $LATEST
2018/05/30 03:06:17 ID [1001653560956858368] Hi, there.
END RequestId: 747fc7fe-5471-13ca-a7ae-8cbeb0335031
REPORT RequestId: 747fc7fe-5471-13ca-a7ae-8cbeb0335031  Duration: 479.12 ms     Billed Duration: 500 ms Memory Size: 128 MB     Max Memory Used: 14 MB
null

The deployment process works in two phases:

1. Package the function and upload it to S3
2. Deploy the function and its supporting infrastructure (referencing the artifact above)

Both of these steps are handled by the sam CLI, which seems to mostly wrap the aws cloudformation CLI.

First, use the package target of the Makefile to upload the binary to S3 and reference it in a newly created packaged.yaml CloudFormation configuration.

$ make package -e AWS_PROFILE=personal
GOOS=linux go build -o main main.go
sam package --s3-bucket neuralyzer-global-config-us-east-1 --template-file template.yaml --output-template-file packaged.yaml                                                    
Uploading to 7001c68762c2fcda61de373e0a30563d  29187040 / 29187040.0  (100.00%)
Successfully packaged artifacts and wrote output template to file packaged.yaml.
Execute the following command to deploy the packaged template
aws cloudformation deploy --template-file packaged.yaml --stack-name <YOUR STACK NAME>

Lastly, use the deploy target of the Makefile to apply the packaged.yaml CloudFormation configuration.

$ make deploy -e AWS_PROFILE=personal 
sam deploy --template-file packaged.yaml --stack-name Neuralyzer --capabilities CAPABILITY_IAM

Waiting for changeset to be created..
Waiting for stack create/update to complete
Successfully created/updated stack - Neuralyzer