Managed Service for Apache Flink
Managed Service for Apache Flink applications: lifecycle, snapshots, logging options and tags, with real Flink jobs in Docker.
Quick start
import boto3
kda = boto3.client("kinesisanalyticsv2", endpoint_url="http://localhost:4566",
region_name="us-east-1",
aws_access_key_id="test", aws_secret_access_key="test")
kda.create_application(ApplicationName="app", RuntimeEnvironment="FLINK-1_20",
ServiceExecutionRole="arn:aws:iam::000000000000:role/flink")
kda.start_application(ApplicationName="app")
Supported operations
15 operations exposed by this service as of MiniStack 1.5.29. Extracted directly from the handler dispatch in the source module.
AddApplicationCloudWatchLoggingOption
CreateApplication
CreateApplicationSnapshot
DeleteApplication
DeleteApplicationCloudWatchLoggingOption
DeleteApplicationSnapshot
DescribeApplication
DescribeApplicationSnapshot
ListApplicationSnapshots
ListTagsForResource
StartApplication
StopApplication
TagResource
UntagResource
UpdateApplication
CloudFormation
The CloudFormation engine provisions these resource types via this service:
AWS::KinesisAnalyticsV2::Application
AWS::KinesisAnalyticsV2::ApplicationCloudWatchLoggingOption
See CloudFormation engine for intrinsic support and lifecycle details.
Known limitations
- Only FLINK-* runtimes; SQL and Studio (Zeppelin) applications are not implemented.
- Jobs run only for FLINK-1_19 and FLINK-1_20 with Docker available; otherwise statuses advance without a running job.
- aws-kinesisanalytics-runtime is not bundled: mount it in /opt/ministack/flink-lib or bundle it in the job jar.
Source
ministack/services/kinesisanalyticsv2.py
Read the source to verify the ops list above — dispatch tables and handler functions are the ground truth.