AWS Lambda

I'm migrating https://www.check-run-reporter.com from EC2 and Aurora Postgres to Lambda and DynamoDB. 

Mostly, I'm doing this for to save costs. When my AWS credits run out, this project will be burning about $200 a month simply because Aurora Serverless is neither serverless nor cheap. More importantly to my customers, this change should also address the latency spikes they keep seeing because Aurora Serverless has been goofy since the AWS outage caused by the Let's Encrypt root certificate expiration this past fall.

Created a very simple AWS serverless application to handle Telegram bot in-coming and out-coming messages with EventBridge.

With this approach message communication can be handled asynchronously outside the connector application just interacting with the exported Event Bridge bus.

Also an username and source IP whitelist is implemented in order to keep the integration as safe as possible.

Repository: daaru00/aws-telegram-bot-connector

Developed cloud-native, serverless architecture using Firecracker, Amazon’s KVM-based virtualization tech. 

Micro-VMs, integrated with the container ecosystem offer a more scalable and efficient compute solution. 

Just created a very simple AWS integration that collect Greengrass Nucleus telemetry data as CloudWatch metrics.

This integration will deploy a Lambda function connect to IoT Core MQTT topic where NucleusEmitter component send telemetry data:

{
  "mqttTopic": "greengrass/MyDeviceName",
  "telemetryPublishIntervalMs": 300000
}

The deployed Lambda function put metrics data to CloudWatch using AWS NodeJS SDK:

  cloudWatch.putMetricData({
    Namespace: METRIC_NAMESPACE,
    MetricData: [{
      MetricName: event.N,
      Unit: event.U,
      Timestamp: event.TS / 1000,
      Dimensions: [{
        Name: 'CoreDevice',
        Value: CORE_DEVICE
      }],
      Value: event.V
    }]
  })

Once this integration is installed it will be possible to add metrics widgets into a dashboard and/or creating alarms in order to monitor Greengrass Core devices health.

Repository: daaru00/greengrass-nucleus-telemetry-data-collector

You can run integration tests from a CloudFormation Custom Resource to enable automatic rollback when your tests fail.

I've written a guide that will get you started called, CDK Integration Testing.

Created a tool for automatically managing access to a restricted AWS environment based on temporarily permitting an engineer's public IP address.