Amazon CloudWatch Synthetics Construct Library

---

End-of-Support

AWS CDK v1 has reached End-of-Support on 2023-06-01. This package is no longer being updated, and users should migrate to AWS CDK v2.

For more information on how to migrate, see the Migrating to AWS CDK v2 guide.

Amazon CloudWatch Synthetics allow you to monitor your application by generating synthetic traffic. The traffic is produced by a canary: a configurable script that runs on a schedule. You configure the canary script to follow the same routes and perform the same actions as a user, which allows you to continually verify your user experience even when you don’t have any traffic on your applications.

Canary

To illustrate how to use a canary, assume your application defines the following endpoint:

% curl "https://api.example.com/user/books/topbook/"
The Hitchhikers Guide to the Galaxy

The below code defines a canary that will hit the books/topbook endpoint every 5 minutes:

canary = synthetics.Canary(self, "MyCanary",
    schedule=synthetics.Schedule.rate(Duration.minutes(5)),
    test=synthetics.Test.custom(
        code=synthetics.Code.from_asset(path.join(__dirname, "canary")),
        handler="index.handler"
    ),
    runtime=synthetics.Runtime.SYNTHETICS_NODEJS_PUPPETEER_3_1,
    environment_variables={
        "stage": "prod"
    }
)

The following is an example of an index.js file which exports the handler function:

const synthetics = require('Synthetics');
const log = require('SyntheticsLogger');

const pageLoadBlueprint = async function () {
  // Configure the stage of the API using environment variables
  const url = `https://api.example.com/${process.env.stage}/user/books/topbook/`;

  const page = await synthetics.getPage();
  const response = await page.goto(url, { waitUntil: 'domcontentloaded', timeout: 30000 });
  // Wait for page to render. Increase or decrease wait time based on endpoint being monitored.
  await page.waitFor(15000);
  // This will take a screenshot that will be included in test output artifacts.
  await synthetics.takeScreenshot('loaded', 'loaded');
  const pageTitle = await page.title();
  log.info('Page title: ' + pageTitle);
  if (response.status() !== 200) {
    throw 'Failed to load page!';
  }
};

exports.handler = async () => {
  return await pageLoadBlueprint();
};

Note: The function must be called handler.

The canary will automatically produce a CloudWatch Dashboard:

UI Screenshot

The Canary code will be executed in a lambda function created by Synthetics on your behalf. The Lambda function includes a custom runtime provided by Synthetics. The provided runtime includes a variety of handy tools such as Puppeteer (for nodejs based one) and Chromium.

To learn more about Synthetics capabilities, check out the docs.

Canary Schedule

You can specify the schedule on which a canary runs by providing a Schedule object to the schedule property.

Configure a run rate of up to 60 minutes with Schedule.rate:

schedule = synthetics.Schedule.rate(Duration.minutes(5))

You can also specify a cron expression with Schedule.cron:

schedule = synthetics.Schedule.cron(
    hour="0,8,16"
)

If you want the canary to run just once upon deployment, you can use Schedule.once().

Configuring the Canary Script

To configure the script the canary executes, use the test property. The test property accepts a Test instance that can be initialized by the Test class static methods. Currently, the only implemented method is Test.custom(), which allows you to bring your own code. In the future, other methods will be added. Test.custom() accepts code and handler properties – both are required by Synthetics to create a lambda function on your behalf.

The synthetics.Code class exposes static methods to bundle your code artifacts:

code.fromInline(code) - specify an inline script.
code.fromAsset(path) - specify a .zip file or a directory in the local filesystem which will be zipped and uploaded to S3 on deployment. See the above Note for directory structure.
code.fromBucket(bucket, key[, objectVersion]) - specify an S3 object that contains the .zip file of your runtime code. See the above Note for directory structure.

Using the Code class static initializers:

# To supply the code from a S3 bucket:
import aws_cdk.aws_s3 as s3
# To supply the code inline:
synthetics.Canary(self, "Inline Canary",
    test=synthetics.Test.custom(
        code=synthetics.Code.from_inline("/* Synthetics handler code */"),
        handler="index.handler"
    ),
    runtime=synthetics.Runtime.SYNTHETICS_NODEJS_PUPPETEER_3_4
)

# To supply the code from your local filesystem:
synthetics.Canary(self, "Asset Canary",
    test=synthetics.Test.custom(
        code=synthetics.Code.from_asset(path.join(__dirname, "canary")),
        handler="index.handler"
    ),
    runtime=synthetics.Runtime.SYNTHETICS_NODEJS_PUPPETEER_3_4
)
bucket = s3.Bucket(self, "Code Bucket")
synthetics.Canary(self, "Bucket Canary",
    test=synthetics.Test.custom(
        code=synthetics.Code.from_bucket(bucket, "canary.zip"),
        handler="index.handler"
    ),
    runtime=synthetics.Runtime.SYNTHETICS_NODEJS_PUPPETEER_3_4
)

Note: Synthetics have a specified folder structure for canaries. For Node scripts supplied via code.fromAsset() or code.fromBucket(), the canary resource requires the following folder structure:
canary/
├── nodejs/
   ├── node_modules/
        ├── <filename>.js
For Python scripts supplied via code.fromAsset() or code.fromBucket(), the canary resource requires the following folder structure:
canary/
├── python/
    ├── <filename>.py
See Synthetics docs.

Running a canary on a VPC

You can specify what VPC a canary executes in. This can allow for monitoring services that may be internal to a specific VPC. To place a canary within a VPC, you can specify the vpc property with the desired VPC to place then canary in. This will automatically attach the appropriate IAM permissions to attach to the VPC. This will also create a Security Group and attach to the default subnets for the VPC unless specified via vpcSubnets and securityGroups.

import aws_cdk.aws_ec2 as ec2

# vpc: ec2.IVpc

synthetics.Canary(self, "Vpc Canary",
    test=synthetics.Test.custom(
        code=synthetics.Code.from_asset(path.join(__dirname, "canary")),
        handler="index.handler"
    ),
    runtime=synthetics.Runtime.SYNTHETICS_NODEJS_PUPPETEER_3_4,
    vpc=vpc
)

Note: By default, the Synthetics runtime needs access to the S3 and CloudWatch APIs, which will fail in a private subnet without internet access enabled (e.g. an isolated subnnet).

Ensure that the Canary is placed in a VPC either with internet connectivity or with VPC Endpoints for S3 and CloudWatch enabled and configured.

See Synthetics VPC docs.

Alarms

You can configure a CloudWatch Alarm on a canary metric. Metrics are emitted by CloudWatch automatically and can be accessed by the following APIs:

canary.metricSuccessPercent() - percentage of successful canary runs over a given time
canary.metricDuration() - how much time each canary run takes, in seconds.
canary.metricFailed() - number of failed canary runs over a given time

Create an alarm that tracks the canary metric:

import aws_cdk.aws_cloudwatch as cloudwatch

# canary: synthetics.Canary

cloudwatch.Alarm(self, "CanaryAlarm",
    metric=canary.metric_success_percent(),
    evaluation_periods=2,
    threshold=90,
    comparison_operator=cloudwatch.ComparisonOperator.LESS_THAN_THRESHOLD
)