Usermanagement

Build status

Code coverage

CI pipeline

Overview

An example of Usermanagement service implemented using Connexion library on top of Flask backed by mysql datastore. Stubs are generated by the swagger-codegen project along with OpenAPI-Spec.

Features

Swagger integration
CI integration with Travis for builds pipeline
APM Montioring integration with Prometheus
Single command service and environment up using Docker compose
Montioring dashboards built with grafana
JWT implementation for login
SHA256 - password hasing
CI pipeline for pushing docker images to registry
Unit test integration with CI pipeline
Code coverage report with CI pipeline
Prodcution deployemnts using chef cookbook

Assumptions

Assuming a ubuntu linux workstation

Requirements

Python 2.7+
Docker version needed 17+

Setting local environment

To setup environment, please execute the following from the projects root directory:

docker-compose up

Check service is up visting

http://localhost:8080/v1/ui/

Login API for ADMIN

Request body for swagger to be provided

{
  "password": "admin",
  "username": "admin"
}

Note: Certain API's need user to login . So for login hit swagger with above payload

Known issue with docker compose / To do

Startup order- removed and not guranteed for service up on latest docker compose version so app restart expected intially till DB is up
To do - Implemnent healthchecks as part of startup on docker compose file for service ordering

APM monitoring prometheus

http://localhost:8080/v1/ui/

Metrics published as part of application with per api call and Status code

flask_request_latency_seconds Flask Request Latency
flask_request_count
process_max_fds
process_open_fds

Sample metrics

flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.005",method="POST"} 1.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.01",method="POST"} 1.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.025",method="POST"} 1.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.05",method="POST"} 1.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.075",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.1",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.25",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.5",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="0.75",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="1.0",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="2.5",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="5.0",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="7.5",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="10.0",method="POST"} 2.0
flask_request_latency_seconds_bucket{endpoint="/v1/login",le="+Inf",method="POST"} 2.0
flask_request_latency_seconds_count{endpoint="/v1/login",method="POST"} 2.0

For more metrics per status code

http://localhost:9000/

Prometheus endpoint

http://localhost:9090/targets

Grafana dashboards

Note: Sigup as admin and dasboard name usermanagemnt-apm. Generate some request for viewing metrics

http://localhost:3000/

Build pipeline architecture

For prod

Code push/PR +-----> Github hooks +-----> Travis CI gets Invoked +-----> Run unit Test +-----> Code coverage (If > Acceptable value) +-----> Docker build(for dev branch) &Create Tag (Only on master build) ----> Push tag to prod via cookbook

TODO/Future imporvements:

Integrate CD
Intergrate alert manager for alerting
Integrate notifications for CI and CD pipeline
Add infra monitoring
Chat ops based deployemnt
Enhance docs

General Note:

The tools mentioned are the ones that I have worked with and have proved to work at large scales. These are a lot of other tools that are available and each has to be picked based on the requirement and scale and other factors

satishkumarsk / mart