Let's go through the following steps to create metric filters using AWS CLI:

1. We first need to prepare our metric transformations, as follows:

[
  {
    "metricName": "HelloCountMetric",
    "metricNamespace": "ServerlessProgrammingCookbook",
    "metricValue": "1"
  }
]

Save this into the metric-transformations.json file.

2. Use the put-metric-filter command to create a metric filter using the metric-transformations.json file, as follows:

aws logs put-metric-filter 
    --log-group-name /aws/lambda/demo-lambda-with-cli 
    --filter-name 'HelloCountFilter' 
    --filter-pattern 'Hello' 
    --metric-transformations file://metric-transformations.json 
    --region us-east-1 
    --profile admin

3. Execute the Lambda a few times, either from the console or from AWS CLI, as follows:

aws lambda invoke 
    --invocation-type RequestResponse 
    --function-name demo-lambda-with-cli 
    --log-type Tail 
    --payload '{"name": "Heartin"}' 
    --region us-east-1 
    --profile admin 
    outputfile.txt

The actual output of the invocation does not matter, as long as it is a success message.

4. Verify the metric from the console by going through the following steps:

1. 1. Log in to AWS Console and go to CloudWatch.
   2. Click on Metrics in the sidebar.
   3. Click on Metrics with no dimensions. We should now see a new metric with the name HelloCountMetric. We can also add HelloCountMetric to the graph or create an alarm for it, as shown in the following screenshot: