AWS

Crawler

Overview

Connect to Amazon Web Services (AWS) to:

• See automatic AWS status updates in your stream
• Get CloudWatch metrics for EC2 hosts without installing the Agent
• Tag your EC2 hosts with EC2-specific information (e.g. availability zone)
• See EC2 scheduled maintenance events in your stream
• Collect CloudWatch metrics and events from many other AWS products

Setup

Setting up the Datadog integration with Amazon Web Services requires configuring role delegation using AWS IAM. To get a better understanding of role delegation, refer to the AWS IAM Best Practices guide.

Role delegation

Choose a method for setting up the necessary AWS role. CloudFormation is recommended.

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Action": [
                "apigateway:GET",
                "autoscaling:Describe*",
                "budgets:ViewBudget",
                "cloudfront:GetDistributionConfig",
                "cloudfront:ListDistributions",
                "cloudtrail:DescribeTrails",
                "cloudtrail:GetTrailStatus",
                "cloudtrail:LookupEvents",
                "cloudwatch:Describe*",
                "cloudwatch:Get*",
                "cloudwatch:List*",
                "codedeploy:List*",
                "codedeploy:BatchGet*",
                "directconnect:Describe*",
                "dynamodb:List*",
                "dynamodb:Describe*",
                "ec2:Describe*",
                "ecs:Describe*",
                "ecs:List*",
                "elasticache:Describe*",
                "elasticache:List*",
                "elasticfilesystem:DescribeFileSystems",
                "elasticfilesystem:DescribeTags",
                "elasticfilesystem:DescribeAccessPoints",
                "elasticloadbalancing:Describe*",
                "elasticmapreduce:List*",
                "elasticmapreduce:Describe*",
                "es:ListTags",
                "es:ListDomainNames",
                "es:DescribeElasticsearchDomains",
                "health:DescribeEvents",
                "health:DescribeEventDetails",
                "health:DescribeAffectedEntities",
                "kinesis:List*",
                "kinesis:Describe*",
                "lambda:AddPermission",
                "lambda:GetPolicy",
                "lambda:List*",
                "lambda:RemovePermission",
                "logs:DeleteSubscriptionFilter",
                "logs:DescribeLogGroups",
                "logs:DescribeLogStreams",
                "logs:DescribeSubscriptionFilters",
                "logs:FilterLogEvents",
                "logs:PutSubscriptionFilter",
                "logs:TestMetricFilter",
                "rds:Describe*",
                "rds:List*",
                "redshift:DescribeClusters",
                "redshift:DescribeLoggingStatus",
                "route53:List*",
                "s3:GetBucketLogging",
                "s3:GetBucketLocation",
                "s3:GetBucketNotification",
                "s3:GetBucketTagging",
                "s3:ListAllMyBuckets",
                "s3:PutBucketNotification",
                "ses:Get*",
                "sns:List*",
                "sns:Publish",
                "sqs:ListQueues",
                "states:ListStateMachines",
                "states:DescribeStateMachine",
                "support:*",
                "tag:GetResources",
                "tag:GetTagKeys",
                "tag:GetTagValues",
                "xray:BatchGetTraces",
                "xray:GetTraceSummaries"
            ],
            "Effect": "Allow",
            "Resource": "*"
        }
    ]
}

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Action": [
                "cloudwatch:Get*",
                "cloudwatch:List*",
                "ec2:Describe*",
                "support:*",
                "tag:GetResources",
                "tag:GetTagKeys",
                "tag:GetTagValues"
            ],
            "Effect": "Allow",
            "Resource": "*"
        }
    ]
}

GovCloud and China

1. Open the AWS integration tile. Click the Install button to install this integration.
2. Select the Access Keys (GovCloud or China Only) tab.
3. Enter your AWS Access Key and AWS Secret Key. Only access and secret keys for GovCloud and China are accepted.
4. Choose the services to collect metrics from on the left side of the dialog.
5. Optionally, add tags to all hosts and metrics.
6. Optionally, monitor a subset of EC2 instances by entering the AWS tags in the textbox to hosts with tag. Note: This also applies to an instance’s attached EBS volumes.
7. Optionally, monitor a subset of Lambdas by entering the AWS tags in the textbox to Lambdas with tag.
8. Click Install Integration.

Log collection

There are two ways of sending AWS service logs to Datadog:

• Kinesis Firehose destination: Use the Datadog destination in your Kinesis Firehose delivery stream to forward logs to Datadog
• Cloudformation: Deploy the Datadog Lambda function which subscribes to S3 buckets or Cloudwatch log group and forward logs to Datadog

We strongly recommend to use the Kinesis Firehose destination when you have to send your logs to multiple destinations. Indeed CloudWatch Log groups can only have one subscriber, but Kinesis streams can have multiple subscribers. By subscribing the Kinesis stream to the log groups, you can have multiple consumers of your log data by subscribing them all to the Kinesis stream.

Data Collected

Metrics

Events

Events from AWS are collected on a per AWS-service basis. Please refer to the documentation of specific AWS services to learn more about the events collected.

Tag

The following tags are collected from AWS integrations. Note: Some tags only display on specific metrics.

Troubleshooting

Discrepancy between your data in CloudWatch and Datadog

There are two important distinctions to be aware of:

1. In AWS for counters, a graph that is set to ‘sum’ ‘1minute’ shows the total number of occurrences in one minute leading up to that point, i.e. the rate per 1 minute. Datadog is displaying the raw data from AWS normalized to per second values, regardless of the time frame selected in AWS. This is why you might see Datadog’s value as lower.
2. Overall, min/max/avg have a different meaning within AWS than in Datadog. In AWS, average latency, minimum latency, and maximum latency are three distinct metrics that AWS collects. When Datadog pulls metrics from AWS CloudWatch, the average latency is received as a single time series per ELB. Within Datadog, when you are selecting ‘min’, ‘max’, or ‘avg’, you are controlling how multiple time series are combined. For example, requesting system.cpu.idle without any filter would return one series for each host that reports that metric and those series need to be combined to be graphed. On the other hand, if you requested system.cpu.idle from a single host, no aggregation would be necessary and switching between average and max would yield the same result.

Metrics delayed

When using the AWS integration, Datadog pulls in your metrics via the CloudWatch API. You may see a slight delay in metrics from AWS due to some constraints that exist for their API.

To begin, the CloudWatch API only offers a metric-by-metric crawl to pull data. The CloudWatch APIs have a rate limit that varies based on the combination of authentication credentials, region, and service. Metrics are made available by AWS dependent on the account level. For example, if you are paying for “detailed metrics” within AWS, they are available more quickly. This level of service for detailed metrics also applies to granularity, with some metrics being available per minute and others per five minutes.

Datadog has the ability to prioritize certain metrics within an account to pull them in faster, depending on the circumstances. Please contact Datadog support for more info.

To obtain metrics with virtually zero delay, install the Datadog Agent on the host. For more information, see Datadog’s blog post Don’t fear the Agent: Agent-based monitoring.

Missing metrics

CloudWatch’s API returns only metrics with data points, so if for instance an ELB has no attached instances, it is expected not to see metrics related to this ELB in Datadog.

Wrong count of aws.elb.healthy_host_count

When the cross-zone load balancing option is enabled on an ELB, all the instances attached to this ELB are considered part of all availability zones (on CloudWatch’s side), so if you have 2 instances in 1a and 3 in ab, the metric displays 5 instances per availability zone. As this can be counter intuitive, we’ve added new metrics, aws.elb.healthy_host_count_deduped and aws.elb.un_healthy_host_count_deduped, that display the count of healthy and unhealthy instances per availability zone, regardless of if this cross-zone load balancing option is enabled or not.

Duplicated hosts when installing the Agent

When installing the Agent on an AWS host, you might see duplicated hosts on the infra page for a few hours if you manually set the hostname in the Agent’s configuration. This second host disappears a few hours later, and won’t affect your billing.