Creating and updating groups of Elastic Beanstalk environments
With the AWS Elastic Beanstalk ComposeEnvironments API, you can
create and update groups of Elastic Beanstalk environments within a single application. Each environment in the group can run a separate component of a service-oriented
architecture application. The Compose Environments
API takes a list of application versions and an optional group name. Elastic Beanstalk
creates an environment for each application version, or, if the environments already exist, deploys the application versions to them.
Create links between Elastic Beanstalk environments to designate one environment as a dependency of another. When you create a group of environments with the
Compose Environments
API, Elastic Beanstalk creates dependent environments only after their dependencies are up and running. For more information on
environment links, see Creating links between Elastic Beanstalk environments.
The Compose Environments
API uses an environment manifest to store configuration details
that are shared by groups of environments. Each component application must have an env.yaml
configuration file in its application
source bundle that specifies the parameters used to create its environment.
Compose Environments
requires the EnvironmentName
and SolutionStack
to be specified in the environment manifest
for each component application.
You can use the Compose Environments
API with the Elastic Beanstalk command line interface (EB CLI), the AWS CLI, or an SDK. See Managing multiple Elastic Beanstalk environments as a group with the EB
CLI for EB CLI instructions.
Using the Compose Environments
API
For example, you could make an application named Media Library
that lets users upload and manage images and videos stored in Amazon Simple Storage Service
(Amazon S3). The application has a front-end environment, front
, that runs a web application that lets users upload and download individual files,
view their library, and initiate batch processing jobs.
Instead of processing the jobs directly, the front-end application adds jobs to an Amazon SQS queue. The second environment, worker
, pulls
jobs from the queue and processes them. worker
uses a G2 instance type that has a high-performance GPU, while front
can run on a
more cost-effective generic instance type.
You would organize the project folder, Media Library
, into separate directories for each component, with each directory containing an
environment definition file (env.yaml
) with the source code for each:
~/workspace/media-library
|-- front
| `-- env.yaml
`-- worker
`-- env.yaml
The following listings show the env.yaml
file for each component application.
~/workspace/media-library/front/env.yaml
EnvironmentName: front+
EnvironmentLinks:
"WORKERQUEUE" : "worker+"
AWSConfigurationTemplateVersion: 1.1.0.0
EnvironmentTier:
Name: WebServer
Type: Standard
SolutionStack: 64bit Amazon Linux 2015.09 v2.0.4 running Java 8
OptionSettings:
aws:autoscaling:launchconfiguration:
InstanceType: m4.large
~/workspace/media-library/worker/env.yaml
EnvironmentName: worker+
AWSConfigurationTemplateVersion: 1.1.0.0
EnvironmentTier:
Name: Worker
Type: SQS/HTTP
SolutionStack: 64bit Amazon Linux 2015.09 v2.0.4 running Java 8
OptionSettings:
aws:autoscaling:launchconfiguration:
InstanceType: g2.2xlarge
After creating an application version for the front-end (front-v1
) and worker
(worker-v1
) application components, you call the Compose Environments
API with the version names. In this example, we use the
AWS CLI to call the API.
# Create application versions for each component:
~$ aws elasticbeanstalk create-application-version --application-name media-library --version-label front-v1 --process --source-bundle S3Bucket="amzn-s3-demo-bucket",S3Key="front-v1.zip"
{
"ApplicationVersion": {
"ApplicationName": "media-library",
"VersionLabel": "front-v1",
"Description": "",
"DateCreated": "2015-11-03T23:01:25.412Z",
"DateUpdated": "2015-11-03T23:01:25.412Z",
"SourceBundle": {
"S3Bucket": "amzn-s3-demo-bucket",
"S3Key": "front-v1.zip"
}
}
}
~$ aws elasticbeanstalk create-application-version --application-name media-library --version-label worker-v1 --process --source-bundle S3Bucket="amzn-s3-demo-bucket",S3Key="worker-v1.zip"
{
"ApplicationVersion": {
"ApplicationName": "media-library",
"VersionLabel": "worker-v1",
"Description": "",
"DateCreated": "2015-11-03T23:01:48.151Z",
"DateUpdated": "2015-11-03T23:01:48.151Z",
"SourceBundle": {
"S3Bucket": "amzn-s3-demo-bucket",
"S3Key": "worker-v1.zip"
}
}
}
# Create environments:
~$ aws elasticbeanstalk compose-environments --application-name media-library --group-name dev --version-labels front-v1 worker-v1
The third call creates two environments, front-dev
and worker-dev
. The API creates the names of the environments by
concatenating the EnvironmentName
specified in the env.yaml
file with the group name
option specified in
the Compose Environments
call, separated by a hyphen. The total length of these two options and the hyphen must not exceed the maximum
allowed environment name length of 23 characters.
The application running in the front-dev
environment can access the name of the Amazon SQS queue attached to the worker-dev
environment by reading the WORKERQUEUE
variable. For more information on environment links, see Creating links between Elastic Beanstalk environments.