DemoAPI

A containerized RESTful HTTP microservice that stores user, group, and membership data.

About

This is a small RESTful HTTP backend, written in Go, that manages simple user, group, and membership records. A user can belong to multiple groups, and a group can have multiple users. The many-to-many relationship is tracked by the membership record.

Dependencies

• Docker

OR

• Go 1.13+
• Sqlite3

How To: with Docker

make up
curl http://localhost:8080/

To cleanup local images:

make clean-docker

How To: from src (without Prometheus and Grafana servers)

cd go/src/demoapi
make runsqlite3-no-generate
curl http://localhost:8080/

Dummy Data

There is also a helper python script that will POST up a number of users and groups and create memberships so that there is dummy data available to poke.

NOTE: The python script depends on the "urllib.requests" and "json" libraries.

make up
python post_dummy_data.py

Example cURLs

• Create a new user

curl -X POST --data-binary '{"first_name": "Matt", "last_name": "F", "userid": "mattf", "groups": ["nasa"]}' http://localhost:8080/users

• Get user data

curl http://localhost:8080/users/{userid}

• Update a specific user's memberships

curl -X PUT --data-binary '{"groups": ["group1", "group2"]}' http://localhost:8080/users/user1

• Delete user and their memberships

curl -X DELETE http://localhost:8080/users/{userid}

• Get all possible users, paged

curl http://localhost:8080/users?quantity=10&offset=10

• Create a new group

curl -X POST --data-binary '{"name": "group1"}' http://localhost:8080/groups

• Get group data

curl http://localhost:8080/groups/{groupname}

• Update a specific group's memberships

curl -X PUT --data-binary '{"userids": ["user1", "user2"]}' http://localhost:8080/groups/group1

• Delete group and its memberships

curl -X DELETE http://localhost:8080/groups/{groupname}

• Get all possible groups, paged

curl http://localhost:8080/groups?quantity=10&offset=10

If you have jq installed:

curl -s http://localhost:8080/users?limit=10&token=10 | jq

Additional features not in the spec

• GET /users Will return paginated user objects with links to the next page
• GET /groups Will return paginated group objects with links to the next page
• GET /metrics Will return Prometheus metrics that can be used by the Grafana server, visible at localhost:3000. (See note below about connecting the Prometheus server as a Grafana datasource)
• An insecure_requests_mode flag that can be set in go/src/demoapi/config.hcl that toggles the need for an Authorization header with each request. For now, the token just needs to be any non-empty string.
• Support for either sqlite or postgres depending on the provided database configuration variables
• The entire project is containerized and stood up with docker-compose.

If the insecure_requests_mode = false configuration is set in config.hcl, then an Authorization token must be provided with each request, like:

curl -H "Authorization: Bearer dummy_token" http://localhost:8080/groups

Using Grafana

make up

This will spin off the main API server with Prometheus metric collection, as well as a Prometheus proxy server and a Grafana server. Once everything is up and running, go to localhost:3000. You should be greeted by the Grafana log in page. On first login, the creds are admin and admin.

Once you're in, go to the datasources tab, select "Prometheus" and enter http://127.0.0.1:9090 as the "URL" and change "Access" to Browser. Then scroll down and test the connection.

At this point, Create a New Dashboard and "Add a new panel". Once here, you should be able to start visualizing metrics collected for the API server.