This is a small example that uses Python and Flask to provision a simple REST api. It uses the Tox build system and Pytest for running tests.
The api will listen for requests on port 8080. Note that this is the port number that the build pipelines will use when defining the ECS tasks and load balancer rules. Changing the api to use a different port will result in incorrect deployment of the resulting container.
It will respond to the following requests:
- /
- /systeminfo
- /systeminfo?detail=all
- /systeminfo?detail=properties
- /systeminfo?detail=environment
Note that the systeminfo requests can return detailed Python system property and environment variable information. Care should to taken to ensure this is only deployed for development purposes.
To enable this Python and PIP need to be installed and configured. The Dockerfile specifies Python version 3.6.
To run locally, execute:
export FLASK_APP=ccsexampleapi2/main.py
flask run --port=8080
This will start the api on port 8080. The run on a different port, for example 18080 simply change the port number that is specified.
You may wish to use a different port when an application and api are being run locally. See the example application repository, CMpExampleApp2.
Now the api is running locally it is possible make requests to it, for example using curl:
curl http://localhost:8080/
will return a JSON object like: {"dateAndTime":"Thursday, 27. September 2018 01:31PM","id":"1c062164-dbb4-40f9-b015-b7798e484ef1"}.
The corresponding example build pipeline project is in the CMpDevEnvironment repository as /terraform/build/api2. The pipeline currently needs to be stored within the CMpDevEnvironment repository because it requires access to various Terraform modules.
The project includes a simple Docker file, this is used by the build pipeline to generate the container image. It is a standard file for a Python project. It must expose port 8080.
To build the Docker image locally execute:
docker build -t <image-name> .
The resulting image can be executed using:
docker run -p 8080:8080 <image-name>