ci-webhook-serverless-example

This example uses the Node.js serverless toolset to create an AWS Lambda function implemented in Python to handle websocket requests from a simulated build system. Requests to the lambda function are logged to a DynamoDB database so that the elapsed time of each step of the build process can be captured. The lambda function reports build started and build complete events to Splunk Infrastucture Monitoring along with metrics for the build and each step ( build_system.build_time_ms and build_system.step_time_ms, respectively)

A sample dashboard is provided illustrating usage of these metrics and annotation via events.

Deploying the example to AWS

1. Install serverless from https://www.serverless.com/. This toolset is used to push lambda functions and supporting components as described via the serverless.yml configuration.

   npm install -g serverless
   cd ci-webhook-handler
   npm init --yes
   npm install --save serverless-python-requirements
   

2. Review the info in serverless.yml. You can change the names of the resources in serverless.yml, but you'll need to make corresponding changes to the lambda function.

3. Create an AWS Credential for serverless to deploy

   There are several options for connecting. See : https://www.serverless.com/framework/docs/providers/aws/guide/credentials/

4. Package the necessary python libraries with the runtime. serverless uses Docker to build a custom runtime by pulling the AWS runtime as a docker image and including dependencies declared in requirements.txt into that image to form a custom runtime. NOTE: THIS MAY TAKE A VERY LONG TIME THE FIRST RUN (if it has to pull base images from AWS Docker repo and run docker build)

   serverless deploy
   

   When that is complete you should see the lambda URL denoted under endpoints: as the sole POST endpoint

5. The configuration in serverless.yml sets up a secret using AWS Secrets Manager to store your Splunk Infrastructure Monitoring credentials. This secret is by default named SignalFx/Ingest. You will need to edit that secret and set the secret values for two keys SignalFxRealm and SignalFxToken.

6. At this point you should have a running lambda function which you can test at the command line via (using the lambda URL which was returned from the serverless deploy command):

   curl -X POST -H "Content-Type: application/json" -d '{ "environment": "env01", "buildId": "abc321", "eventType": "start_build", "buildStep": "start", "status": "success" }' <lambda_url>
   

   (wait a few seconds)

   curl -X POST -H "Content-Type: application/json" -d '{ "environment": "env01", "buildId": "abc321", "eventType": "build_step", "buildStep": "step1", "status": "success" }' <lambda_url>
   

   (wait a few seconds)

   curl -X POST -H "Content-Type: application/json" -d '{ "environment": "env01", "buildId": "abc321", "eventType": "build_complete", "buildStep": "finito", "status": "success" }' <lambda_url>
   

7. You should now be able to log into Splunk Infrastructure Monitoring and create a new chart showing metrics build_system.build_time_ms and build_system.step_time_ms.

8. You can generate some fake data with:

   python3 generate-test-events.py <lambda_url> <environment>
   

9. You can import the custom dashboard dashboard_Build System Demo.json. You should see metrics and events depicted in that dashboard.