Monitor Amazon Web Services (AWS) with Beats

In this tutorial, you’ll learn how to monitor your AWS infrastructure using Elastic Observability: Logs and Infrastructure metrics.

You’ll learn how to:

Create a deployment using our hosted Elasticsearch Service on Elastic Cloud. The deployment includes an Elasticsearch cluster for storing and searching your data, and Kibana for visualizing and managing your data.

With this tutorial, we assume that your logs and your infrastructure data are already shipped to CloudWatch. We are going to show you how you can stream your data from CloudWatch to Elasticsearch. If you don’t know how to put your AWS logs and infrastructure data in CloudWatch, Amazon provides a lot of documentation around this specific topic:

To centralize your logs in Elasticsearch, you need to have an S3 bucket. Filebeat, the agent you’ll use to collect logs, has an input for S3.

In the AWS S3 console, click on Create bucket. Give the bucket a name and specify the region in which you want it deployed.

You should now have an S3 bucket in which you can export your logs, but you will also need an SQS queue. To avoid significant lagging with polling all log files from each S3 bucket, we will use Amazon Simple Queue Service (SQS). This will provide us with an Amazon S3 notification when a new S3 object is created. The Filebeat S3 input checks SQS for new messages regarding the new object created in S3 and uses the information in these messages to retrieve logs from S3 buckets. With this setup, periodic polling from each S3 bucket is not needed. Instead, the Filebeat S3 input guarantees near real-time data collection from S3 buckets with both speed and reliability.

Create an SQS queue and configure our S3 bucket to send a message to the SQS queue whenever new logs are present in the S3 bucket. Go to the SQS console

Create a standard SQS queue and edit the access policy by using a JSON object to define an advanced access policy:

{
  "Version": "2012-10-17",
  "Id": "example-ID",
  "Statement": [
    {
      "Sid": "example-statement-ID",
      "Effect": "Allow",
      "Principal": {
        "AWS": "*"
      },
      "Action": "SQS:SendMessage",
      "Resource": "<sqs-arn>",
      "Condition": {
        "StringEquals": {
          "aws:SourceAccount": "<source-account>"
        },
        "ArnLike": {
          "aws:SourceArn": "<s3-bucket-arn>"
        }
      }
    }
  ]
}

Now that your queue is created, go to the properties of the S3 bucket you created and click Create event notification.

Specify that you want to send a notification on every object creation event.

Set the destination as the SQS queue you just created.

To monitor AWS using the Elastic Stack, you need two main components: an Elastic deployment to store and analyze the data and an agent to collect and ship the data.

Download and install Filebeat.

Filebeat comes with predefined assets for parsing, indexing, and visualizing your data. Run the following command to load these assets. It may take a few minutes.

./filebeat setup -e -E 'cloud.id=YOUR_DEPLOYMENT_CLOUD_ID' -E 'cloud.auth=elastic:YOUR_SUPER_SECRET_PASS' 

Next, you are going to configure Filebeat output to Elasticsearch Service.

Now that the output is working, you can set up the Filebeat AWS module which will automatically create the AWS input. This module checks SQS for new messages regarding the new object created in the S3 bucket and uses the information in these messages to retrieve logs from S3 buckets. With this setup, periodic polling from each S3 bucket is not needed.

There are many different filesets available: cloudtrail, vpcflow, ec2, cloudwatch, elb and s3access. In this tutorial, we are going to show you a few examples using the ec2 and the s3access filesets.

The ec2 fileset is used to ship and process logs stored in CloudWatch, and export them to an S3 bucket. The s3access fileset is used when S3 access logs need to be collected. It provides detailed records for the requests that are made to a bucket. Server access logs are useful for many applications. For example, access log information can be useful in security and access audits. It can also help you learn about your customer base and understand your Amazon S3 bill.

Let’s enable the AWS module in Filebeat.

./filebeat modules enable aws

Edit the modules.d/aws.yml file with the following configurations.

- module: aws
  cloudtrail:
    enabled: false
  cloudwatch:
    enabled: false
  ec2:
    enabled: true 
    var.credential_profile_name: fb-aws 
    var.queue_url: https://sqs.eu-central-1.amazonaws.com/836370109380/howtoguide-tutorial 
  elb:
    enabled: false
  s3access:
    enabled: false
  vpcflow:
    enabled: false

Make sure that the AWS user used to collect the logs from S3 has at least the following permissions attached to it:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Action": [
              "s3:GetObject",
              "sqs:ReceiveMessage",
              "sqs:ChangeMessageVisibility",
              "sqs:DeleteMessage"
            ],
            "Effect": "Allow",
            "Resource": "*"
        }
    ]
}

You can now upload your logs to the S3 bucket. If you are using CloudWatch, make sure to edit the policy of your bucket as shown in step 3 of the AWS user guide. This will help you avoid permissions issues.

Start Filebeat to collect the logs.

./filebeat -e

Here’s what we’ve achieved so far:

Now, let’s configure the s3access fileset. The goal here is to be able to monitor how people access the bucket we created. To do this, we’ll create another bucket and another queue. The new architecture will look like this:

Create a new S3 bucket and SQS queue. Ensure that the event notifications on the new bucket are enabled, and that it’s sending notifications to the new queue.

Now go back to the first bucket, and go to Properties > Server access logging. Specify that you want to ship the access logs to the bucket you most recently created.

Copy the URL of the queue you created. Edit the `modules.d/aws.yml`file with the following configurations.

- module: aws
  cloudtrail:
    enabled: false
  cloudwatch:
    enabled: false
  ec2:
    enabled: true 
    var.credential_profile_name: fb-aws 
    var.queue_url: https://sqs.eu-central-1.amazonaws.com/836370109380/howtoguide-tutorial 
  elb:
    enabled: false
  s3access:
    enabled: true 
    var.credential_profile_name: fb-aws 
    var.queue_url: https://sqs.eu-central-1.amazonaws.com/836370109380/access-log 
  vpcflow:
    enabled: false

Once you have edited the config file, you need to restart Filebeat. To stop Filebeat, you can press CTRL + C in the terminal. Now let’s restart Filebeat by running the following command:

./filebeat -e

Now that the logs are being shipped to Elasticsearch we can visualize them in Kibana. To see the raw logs, open the main menu in Kibana, then click Logs:

The filesets we used in the previous steps also come with pre-built dashboards that you can use to visualize the data. In Kibana, open the main menu and click Dashboard. Search for S3 and select the dashboard called: [Filebeat AWS] S3 Server Access Log Overview:

This gives you an overview of how your S3 buckets are being accessed.

To monitor your AWS infrastructure you will need to first make sure your infrastructure data are being shipped to CloudWatch. To ship the data to Elasticsearch we are going to use the AWS module from Metricbeat. This module periodically fetches monitoring metrics from AWS CloudWatch using GetMetricData API for AWS services.

In a new terminal window, run the following commands.

Download and install Metricbeat.

Metricbeat comes with predefined assets for parsing, indexing, and visualizing your data. Run the following command to load these assets. It may take a few minutes.

./metricbeat setup -e -E 'cloud.id=YOUR_DEPLOYMENT_CLOUD_ID' -E 'cloud.auth=elastic:YOUR_SUPER_SECRET_PASS' 

Next, you are going to configure Metricbeat output to Elasticsearch Service.

Now that the output is working, you are going to set up the AWS module.

To collect metrics from your AWS infrastructure, we’ll use the Metricbeat AWS module. This module contains many metricsets: billing, cloudwatch, dynamodb, ebs, ec2, elb, lambda, and many more. Each metricset is created to help you stream and process your data. In this tutorial, we’re going to show you a few examples using the ec2 and the billing metricsets.

Make sure that the AWS user used to collect the metrics from CloudWatch has at least the following permissions attached to it:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Action": [
                "ec2:DescribeInstances",
                "ec2:DescribeRegions",
                "cloudwatch:GetMetricData",
                "cloudwatch:ListMetrics",
                "sts:GetCallerIdentity",
                "iam:ListAccountAliases",
                "tag:getResources",
                "ce:GetCostAndUsage"
            ],
            "Effect": "Allow",
            "Resource": "*"
        }
    ]
}

You can now start Metricbeat:

./metricbeat -e

Now that the metrics are being streamed to Elasticsearch we can visualize them in Kibana. In Kibana, open the main menu and click Infrastructure. Make sure to show the AWS source and the EC2 Instances:

The metricsets we used in the previous steps also comes with pre-built dashboard that you can use to visualize the data. In Kibana, open the main menu and click Dashboard. Search for EC2 and select the dashboard called: [Metricbeat AWS] EC2 Overview:

If you want to track your billings on AWS, you can also check the [Metricbeat AWS] Billing Overview dashboard: