It is a Django template for django-admin startproject command that provides you containerized (docker) sets of components cooperating together. The goal is to maximize the time for coding having all devops/pipeline/testing stuff ready to improve. This project aims to be python centric, although there are some tools where good python replacement does not exist. Yet.
Currently there are couple of things set up and ready to be used and improved:
- Central logs aggregation with ELK
- Access to all containers logs and shells via simple
makecommands - Performance testing with Locust.io supported by SBE (via Behave)
- Monitoring of all components with Glances + Graphite + Grafana
- SBE/BDD testing with Behave
- Docker console with Portainer
- Building docs for your services with Sphinx
- CI/CD pipline fired at local commits with Buildbot
- Plugged in service for you app based on Django and supported by Gunicorn, Nginx, Postgresql
In the next releases:
- NoSql DB
- Key-value store
- Cache
- Persistent queues
- Alerting
- Load balancing
- Service discovery
- API gateway
- Frontend with SSR (Vue.js)
- Event sourcing approach built in
You need Linux machine (tested on Ubuntu 18.04) with docker engine, make, virtualenv with Django and docker-compose. If you want to set up CI-CD env locally then you need docker-machine and Git as well.
Dependencies:
Docker >= 19.03
Docker-compose >= 1.27
Django >= 3.1
Git >= 2.25
How to install docker see here.
Configuration for docker-compose and Django:
python3 -m venv .venv
source .venv/bin/activate
python -m pip install Django==3.1 docker-compose==1.27 GitPython==3.1 requests==2.20
python -m pip install --upgrade pip
To create source code for your service based on this template you need to run:
django-admin startproject \
--template=https://github.com/paterit/django-microservice-template/archive/master.zip \
--extension=py,rst,yml,sh,md,conf,feature \
--name=Makefile,locustfile.py,performance-testing.json,perf.py,glances-graphite.json,Dockerfile-perf,Dockerfile-glances,Dockerfile-grafana,Dockerfile-docs,Dockerfile-web,Dockerfile-db,Dockerfile-data,Dockerfile-https,Dockerfile-testing,Dockerfile,master.cfg,db.env,docker-machine.docker.env,post-commit \
yourservice
Due to docker-machine limits on naming machines don't use "_" (underscore) sign when naming your project. Mainly due to resource-hungry ELK stack you should have at least 4GB of RAM on your dev machine.
In order to have ElasticSearch (part of ELK stack) working on your machine you have to run:
sudo sysctl -w vm.max_map_count=262144
Now you can download all needed docker images and build your containers just by typing make in yourservice directory:
cd yourservice
make
To check if it runs properly verify if new containers are running by typing:
docker ps
You should see among running containers some with names like :
yourservice-db - PostgeSQL
yourservice-web - Django application with Gunicorn
yourservice-https - Nginx
yourservice-testing - Behave, Selenium, PhantomJS
yourservice-logs - ELK stack
yourservice-logspout - Logspout - log forwarder from Docker to Logstash
yourservice-docker-console - Portainer - web docker console
yourservice-monitoring-agent - Glances - monitoring agent
yourservice-monitoring-server - Graphit+Grafana - monitoring server
yourservice-https-logs - Logs for Nginx
yourservice-web-logs - Logs for Django and Gunicorn
If they are up and running you should be able to see admin panel (user: admin, password: admin)
To read how it can be used go to the docs.
To see any other useful links go to this page in docs.
All of it creates a bunch of images and containers. But don't worry. It can be easily and safely cleaned up by running:
make clean-all
To create a virtual machine with local CI/CD machinery you need to install docker-machine and pv command ( sudo apt-get install pv ).
The following commands will set up local docker-machine and local docker containers with buildbot which will allow you to run and test your code within docker-machine. Start with:
cd yourservice
make dev-docker-machine
To check if it runs propelly verify if new containers are running by typing:
docker ps
You should see among running containers some with names like :
yourservice-cicd-worker - Buildbot worker
yourservice-cicd-master - Buidbot master
yourservice-cicd-db - Buildbot database
Verify if docker-machine is running by typing:
docker-machine ls
You should see among others machines one with the name:
yourservice-cicd
Now you are able to use Buildbot through its web interface. There are prepared builders that allow to build, run and test all containers in docker-machine. For the first time, you have to run at least once "Full rebuild" builder. While running it for the first time couple GBs of data will be downloaded so it may take a while. All base images for docker need to be downloaded to docker machine (just to name a few: Python, PostgreSQL, ELK, Nginx). This "Full rebuild" builder should be already started (this is the last step in setting up local CICD env). From now on whenever you commit any change locally to your project there will be message send to CICD via post-commit hook in your local git.
Now using IP generated for your docker-machine machine (in my case it is 192.168.99.100 - you can verify it by reading the output from docker-machine ls ) you can start using your services. Django admin panel (user: admin, password: admin) To read how it can be further used go to docs. To see any other useful links go to this page in docs.
Whenever you do changes in your code, when you run any builders in Buildbot the fresh copy of your sources will be copied to Buildbot worker and tested.
To clean CICD containers from run:
make clean-cicd
A bunch of images and containers exists in docker-machine. It can be easily and safely cleaned up by running (it will neither stop or remove docker-machine itself):
make clean-docker-machine
To stop or remove docker machine type either:
docker-machine stop yourservice-cicd
or:
docker-machine rm yourservice-cicd
The following commands will set up local containers with buildbot which will allow you to run and test your code within remote docker host. Environment variables to connect CICD to remote docker host should be defined in file remote.docker.env. In order to have ElasticSearch (part of ELK stack) working on your remote docker machine you have to run on its host OS:
sudo sysctl -w vm.max_map_count=262144
Then continue on your local machine:
cd yourservice
make remote
To check if it runs propelly verify if new containers are running by typing:
docker ps
You should see among running containers some with names like :
yourservice-cicd-worker - Buildbot worker
yourservice-cicd-master - Buidbot master
yourservice-cicd-db - Buildbot database
Now you are able to use Buildbot through its web interface. There are prepared builders that allow to build, run and test all containers in docker-machine. For the first time, you have to run at least once "Full rebuild" builder. While running it for the first time couple GBs of data will be downloaded so it may take a while. All base images for docker need to be downloaded to docker machine (just to name a few: Python, PostgreSQL, ELK, Nginx). This "Full rebuild" builder should be already started (this is the last step in setting up local CICD env). From now on whenever you commit any change locally to your project there will be message send to CICD via post-commit hook in Git.
Now with remote docker IP you can start using your services. Django admin panel (user: admin, password: admin) To read how it can be further used go to docs. To see any other useful links go to this page in docs.
Whenever you do changes in your code, when you run any builders in Buildbot the fresh copy of your sources will be copied to Buildbot worker and tested.
To clean CICD containers from run:
make clean-cicd
A bunch of images and containers exists on the remote docker host. It can be easily and safely cleaned up by running:
make clean-remote-docker-images