AWS CloudWatch
Hosted Graphite provides an Amazon AWS CloudWatch integration available in the Add-Ons page of your account. This add-on syncs the metrics from the specified AWS services/regions into your HG account.
To connect to your CloudWatch account, you need to set up Identity and Access Management (IAM) access keys in your AWS Account, with the appropriate permissions to allow Hosted Graphite to connect and collect your metrics.
Contents
Let’s create a policy that we will later attach to the user.
In the IAM Dashboard, click on Policies (on the left) or Customer Managed Policies in under IAM Resources, then click on Create Policy.
Switch to the JSON tab, replace the existing text with the policy provided below, and click ‘Review policy‘.
Give the policy a name (Eg: HG_policy) and a description (optional).
Save the policy by clicking ‘Create Policy‘.
With the policy set up, we’ll create a user to attach the policy to. We’ll use the Access Key/Secret Key tokens to give Hosted Graphite permission to import CloudWatch metric data.
In the IAM Console, click on ‘Users‘ on the left or under IAM Resources. Then select ‘Add User‘.
Give the user a name (Eg: HG_addon) and ‘Programmatic access‘ for the Access type.
Click ‘Next: Permissions‘ and click the ‘Attach existing policies directly‘ button at the top.
Search for the policy we just created and attach it by clicking the corresponding checkbox, then press ‘Next: Review‘.
Click on ‘Create User‘ and copy the Access Key and Secret Key into the relevant fields on the add-on set-up page here. Optionally, you can download a copy of this account information, but keep it safe as this information can be sensitive.
To enable the CloudWatch add-on, go to the add-ons page in your Hosted Graphite account and choose the option for Amazon AWS CloudWatch.
You will be presented with the following page, from which you can create, edit or delete AWS configs or create, edit or delete tags:
If you click AWS Tags you will be presented with the following screen which allows you to edit the tags. Each tag has a unique name (shown on the left) and one or more values. Values are separated by a comma. If for example you enter “hello, world” for tag “Tag_Next” then the CloudWatch Add-On will match any tag called “Tag_Next” that has either “hello” or “world” as its value.
If you click Cancel and go back to the previous screen and click Account Names, you can see this screen from which you can edit a particular AWS configuration.
AWS Access Key and AWS Secret Access Key
These are the keys you saved when you set up your IAM user.
AWS Regions
Choose the regions containing the services you wish to monitor. You must choose at least one region.
AWS Services
Choose the services which will be queried for metrics.
EC2 Instance Aliasing
Choose this if you would like Hosted Graphite to import your EC2 instance metrics using an instance’s name instead of its ID. For instance aliasing to work, it is required that your instances have a “Name” tag defined whose value will be used in place of the instance id. Please note that enabling this feature will create new metrics in your account for each of your EC2 instances that have the “Name” tag. The old instance id based metrics will remain dormant until they expire. Name tags containing spaces are currently not supported and will be dropped
EMR Cluster Aliasing
Select this if you would like Hosted Graphite to import your EMR cluster metrics using a cluster’s name instead of its ID. For cluster aliasing to work, it is required that your clusters have a “Name” tag defined whose value will be used in place of the cluster id. Please note that enabling this feature will create new metrics in your account for each of your EMR clusters that have the “Name” tag. The old cluster id based metrics will remain dormant until they expire. Name tags containing spaces are currently not supported and will be dropped
Service tagging for this AWS account
Choose the services that you would like to enable tagged imports for. This will only import metrics from tagged resources for those services. Once you enable this per account, you can type in the tag’s “key”: “value” pairs that you would like to be imported. These key values must also be present in your AWS resources to be imported successfully. For more information on adding these tags go to the AWS docs.
When you click on ‘Save‘, some basic checks will be performed on your keys if successful, your configuration will be saved. If you have at least one service chosen, the Cloudwatch add-on will be enabled.
Go to the add-ons page in your Hosted Graphite account, and choose the option for Amazon AWS CloudWatch.
Click the Delete button and the CloudWatch add-on will be disabled for that account.
If you use multiple AWS Access Keys on your Hosted Graphite account for different AWS accounts, projects or environments, you can keep the metrics separated by assigning Account Names to your Access Keys.
Under the list of Access Keys on your HG account, there’s a button to access the Account Names interface. There you can assign names to your Access Keys, which will then become part of the metric name for all metrics retrieved using that key. In the example below, the Access Key was assigned the name ‘test’, so metrics retrieved through that access key will follow the naming structure ‘aws.test.[service]*’
If you were previously not using an Account Name and have recently added one, your automatically generated AWS Dashboards will no longer map to the correct metrics. You can tick the option on the Account Names screen (above) to generate new dashboards which use the new Account Names.
Amazon Route53 metrics are only received if you enable US East (N. Virginia) as the current region. These metrics are not available from any other region.
While we attempt to minimize the number of API calls that may incur Amazon charges, Hosted Graphite disclaims responsibility for potential costs incurred by the use of this add-on.
Our add-on performs read-only requests to the CloudWatch API. Should the provided AWS Access Keys grant greater privileges than what our specified IAM Policy defines, responsibility for any activity performed using those keys lies with the customer.
Amazon AWS offers the first one million API requests at no charge. In excess of that, Amazon will charge $0.01 per 1000 requests. See the CloudWatch Pricing page for more information on Amazon pricing.
We make calls to every 10min fetch your metrics from AWS, but feel free to reach out to support@hostedgraphite.com if you would like us to configure a custom pull interval for you.
We aim to make as few requests as possible to fetch your metrics. If you have many instances, or are monitoring many services, you will likely exceed this boundary. If this is the case, these charges will likely be negligible in comparison to what Amazon charges just to have monitoring enabled for those services.
This can be located and generated from within the Dashboard Library, and offers a great starting point for basic AWS monitoring!
This overview dashboard assumes that you have enabled Account Names, which can be configured in Add-Ons => AWS.
The AWS metrics for each service are mapped to Hosted Graphite metric names as follows:
service - A short token representing the service, e.g. ec2, or rds.
region - The AWS region, e.g. us-east-1.
grouping - A short token representing the grouping (‘dimension’ in AWS speak) for the metric, e.g. inst for InstanceId.
id - The identifier for the service/instance, e.g. Instance id or Name tag for EC2.
metricname - The AWS metric name is directly used, e.g. CPUUtilization.
| AWS Service | AWS ‘Dimension’ | HG metric name |
|---|---|---|
Elastic Compute Cloud (EC2) | InstanceId | aws.ec2.{region}.inst.{id}.CPUUtilization |
Elastic Block Store (EBS) | VolumeId | aws.ebs.{region}.vol.{id}.VolumeWriteBytes |
Relational Database Service (RDS) | DBInstanceIdentifier | aws.rds.{region}.inst.{id}.CPUUtilization |
Elastic MapReduce (EMR) | JobFlowId | aws.emr.{region}.job_flow_id.{id}.CPUUtilization |
So, for example, the CPUUtilization metric for the ‘i-abcd1234‘ EC2 instance in Virginia will be imported as ‘aws.ec2.us-east-1.inst.i-abcd1234.CPUUtilization‘. You can find the mapping for each individual AWS service we support below:
If you are having trouble querying Hosted Graphite metrics that have been imported from your AWS services perhaps you are not entering the metric query correctly.
The format of these metrics changes per service and has multiple formats per service. These formats depend on ‘Dimensions’ in the AWS metric or in some cases if the AWS metric came from an AWS ARN (Amazon Resource Name).
For further information on AWS Dimensions and AWS ARNs please consult their corresponding AWS documentation.
Here is a list of how each individual service’s metrics are formatted in Hosted Graphite:
| AWS Service | AWS ‘Dimensions’ | HG metric names |
|---|---|---|
API Gateway | ApiName, Method, Resource, Stage | aws.apigateway.{region}.apiname.{ApiName}.method.{Method}.resource.{Resource}.stage.{Stage}.{name} |
ApiName, Stage | aws.apigateway.{region}.apiname.{ApiName}.stage.{Stage}.{name} | |
ApiName | aws.apigateway.{region}.apiname.{ApiName}.{name} | |
Application Elastic Load Balancing | LoadBalancer, AvailabilityZone, TargetGroup | aws.alb.{region}.inst.{LoadBalancer}.{AvailabilityZone}.{TargetGroup}.{name} |
LoadBalancer | aws.alb.{region}.inst.{LoadBalancer}.{name} | |
arn_format: arn:aws:elasticloadbalancing:{region}:{account-id}:loadbalancer/{loadbalancername} | ||
AutoScaling Groups | AutoScalingGroupName | aws.autoscaling.{AutoScalingGroupName}.{name} |
CloudFront | DistributionId | aws.cloudfront.distribution_id.{DistributionId}.{name} |
CloudWatch Synthetics | CanaryName | aws.cloudwatchsynthetics.{region}.canaryname.{CanaryName}.{name} |
Direct Connect | ConnectionId, VirtuaInterfaceId | aws.directconnect.{region}.connection_id.{ConnectionId}.virtual_interface_id. {VirtualInterfaceId}.{name} |
ConnectionId, OpticalLaneNumber | aws.directconnect.{region}.connection_id.{ConnectionId}.optical_lane_number.{OpticalLaneNumber}.{name} | |
DynamoDB | TableName, Operation, GlobalSecondaryIndexName | aws.dynamodb.{region}.table.{TableName}.operation.{Operation}.secondary_index. {GlobalSecondaryIndexName}.{name} |
Elastic Block Store | VolumeId | aws.ebs.{region}.vol.{VolumeId}.{name} |
Elastic Compute Cloud | InstanceId | aws.ec2.{region}.inst.{InstanceId}.{name} |
EC2 Spotfleet | AvailabilityZone, FleetRequestId, InstanceType | aws.spotfleet.{AvailabilityZone}.{FleetRequestId}.{InstanceType}.{name} |
FleetRequestId, InstanceType | aws.spotfleet.{FleetRequestId}.{InstanceType}.{name} | |
FleetRequestId, AvailabilityZone | aws.spotfleet.{AvailabilityZone}.{FleetRequestId}.{name} | |
FleetRequestId | aws.spotfleet.{FleetRequestId}.{name} | |
ECS | ClusterName, ServiceName | aws.ecs.{region}.cluster.{ClusterName}.service.{ServiceName}.{name} |
ClusterName | aws.ecs.{region}.cluster.{ClusterName}.{name} | |
ECS Container Insights | ClusterName | aws.ecs_containerinsights.{region}.cluster.{ClusterName}.{name} |
ClusterName, ServiceName | aws.ecs_containerinsights.{region}.cluster.{ClusterName}.service.{ServiceName}.{name} | |
ClusterName, TaskDefinitionFamily | aws.ecs_containerinsights.{region}.cluster.{ClusterName}.task_definition_family.{TaskDefinitionFamily}.{name} | |
Elastic File System | FileSystemId | aws.efs.{region}.FileSystemId.{FileSystemId}.{name} |
Elastic Load Balancing | LoadBalancerName, AvailabilityZone | aws.elb.{region}.inst.{LoadBalancerName}.{AvailabilityZone}.{name} |
LoadBalancerName | aws.elb.{region}.inst.{LoadBalancerName}.{name} | |
Elasticsearch Service | DomainName, NodeId, ClientId | aws.es.{region}.{DomainName}.{NodeId}.{ClientId}.{name} |
DomainName, ClientId | aws.es.{region}.{DomainName}.{ClientId}.{name} | |
ElastiCache | CacheClusterId, CacheNodeId | aws.elasticache.{region}.cache_cluster_id.{CacheClusterId}.cache_node_id.{CacheNodeId}.{name} |
CacheClusterId | aws.elasticache.{region}.cache_cluster_id.{CacheClusterId}.{name} | |
arn_format: arn:aws:elasticache:{region}:{account_id}:cluster:{instance_name} | ||
Elastic MapReduce | JobFlowId | aws.emr.{region}.job_flow_id.{JobFlowId}.{name} |
Kinesis Firehose | DeliveryStreamName | aws.firehose.{region}.{DeliveryStreamName}.{name} |
Kinesis Streams | StreamName | aws.kinesis.{region}.stream_name.{StreamName}.{name} |
Lambda Functions | FunctionName | aws.lambda.{region}.function.{FunctionName}.{name} |
Logs | LogGroupName | aws.logs.{region}.{LogGroupName}.{name |
LogGroupName, DestinationType, FilterName | aws.logs.{region}.{LogGroupName}.destination_type.{DestinationType}.filter_name.{FilterName}.{name} | |
Resource, Service, Type, Class | aws.logs.{region}.resource.{Resource}.service.{Service}.type.{Type}.class.{Class}.{name} | |
MemoryDB | ClusterName, NodeName | aws.memorydb.{region}.cluster.{ClusterName}.node.{NodeName}.{name} |
Role, ReplicationGroupId | aws.memorydb.{region}.role.{Role}.replication_group_id.{ReplicationGroupId}.{name} | |
MQ | Broker, Topic | aws.mq.{region}.{Broker}.{Topic}.{name} |
Broker, Queue | aws.mq.{region}.{Broker}.{Queue}.{name} | |
Broker | aws.mq.{region}.{Broker}.{name} | |
Relational Database Service | DBInstanceIdentifier | aws.rds.{region}.inst.{DBInstanceIdentifier}.{name} |
arn_format: arn:aws:rds:{region}:{account_id}:db:{instance_name} | ||
Redshift | NodeID, ClusterIdentifier | aws.redshift.{region}.cluster.{ClusterIdentifier}.node.{NodeID}.{name} |
ClusterIdentifier | aws.redshift.{region}.cluster.{ClusterIdentifier}.{name} | |
NodeID | aws.redshift.{region}.node.{NodeID}.{name} | |
Route 53 | HealthCheckId | aws.route53.{region}.health_check_id.{HealthCheckId}.{name} |
S3 (Simple Storage Service) | BucketName, StorageType | aws.s3.{region}.bucket.{BucketName}.type.{StorageType}.{name} |
Simple E-mail Service | aws.ses.{region}.{name} | |
Simple Notification Service | TopicName | aws.sns.{region}.{TopicName}.{name} |
SMSType, Country | aws.sns.{region}.{SMSType}.{Country}.{name} | |
aws.sns.{region}.{name} | ||
Simple Queue Service | QueueName | aws.sqs.{region}.queue.{QueueName}.{name} |
Last updated
