UserSrv a testing ground for Microservice experiments

This project evolved from a simple testing ground to a more full-fletched time tracker solution. My intend is to make this a viable tool for all kind of time tracking/reporting - i.E. for freelance workers working on different projects and trying to create reports out of ther work based on hours.

Tech used

golang as language for backend
ReactJS/TypeScript/Node as frontend system
GraphQL as API service (frontend2backend)
GRPC for calling backend2backend
zipkin/jaeger as tracing app
etcd for discovery (not used currently)
docker-compose for bringing up everything together on a dev system.
Kubernetes as a system to orchestrate and make it available. Using a Minikube instance with apache as a router on a LinuxVM for now.

Some libraries to mention:

Reactjs, materialui, create-react-app, typescript
ent as entity framework https://github.com/facebookincubator/ent
graphql backend github.com/graph-gophers/graphql-go, code generator used to precreate some basic model classes (github.com/DealTap/graphql-gen-go).
opentracing (grpc, zipkin) to support tracing functionality.

Services (excluding external)

Frontend

Little React project using create-react-app with Typescript
material-ui as UI/UX lib.
nginx hosting page
nginx proxy graphql/rest requests to API service
Authorization uses Microsoft, Google OAuth2 services. Token will be evaluated in backend. New JWT Token will be created with clains.

API Service

handling of login/auth via JWT token
providing GraphQL service/discovery
Forwarding requests to other GRPC based services.
JWT token will be retrieved from cookie/authorization handler and forwarded to grpc based services.
Each service could decide which data it will give based on the decoded JWT token.
GraphQL gateway could also filter data based on JWT.

User Service

sample service storing username, email etc. is serialized to ENT orm to Postgresql
interfaced via GRPC protobuf services.
planned Teams support to make it easier to work on bigger projects and keep track on each team member.

Timer Service

Handling timer entries. is serialized to ENT orm to Postgresql
interfaced via GRPC protobuf services.

Client Service

Handling clients for one user. is serialized to ENT orm to Postgresql
Planned on extending that so that multiple user could work on one client and group users to teams.
interfaced via GRPC protobuf services.

Project Service

Handling projects for one user. is serialized to ENT orm to Postgresql
Planned on extending that so that multiple user could work on one project via a team teams.
interfaced via GRPC protobuf services.

Makefile for handling all the stuff

make rungql runs the API GraphQL service
make runuser/runclient/runtimer/runproject runs corresponding service
make runfrontend runs the frontendservice
make compose-build builds all the docker images for usage in docker-compose
make compose-up runs the composed app completely
make docker builds all images and push them to docker hub
make kapply applies all configs to your kubernetes cluster. Be sure to deploy some services/deployments like jaeger seperately. Also all the secrets need to me manually changed.
make kreapply restarts the deployment to make sure all versions are up to date.

Currently all docker images will be tagged with "latest". Github actions will be triggered to build everything via docker - but not pushed. On my linux staging machine a github hook is listending and rebuilding everything including push and redeploy. So the app will be updateded in a few minutes automatically.

Plans on using helm. This might help to deploy in a more secure way as using versioned image names - and be able to rollback.

some tooling and instructions (basically to remember stuff)

Generating proto models

sh protogen.sh

Testing the gRPC interface

gRPC Testing App

brew cask install bloomrpc

Download via https://github.com/uw-labs/bloomrpc

Service discovery with etcd

etcdctl put my-service/1.2.3.4 '{"Addr":"localhost:10000","Metadata":"..."}'

fixing some strange dependency (probably not necessary anymore)

rm -r $GOPATH/src/go.etcd.io/etcd/vendor/golang.org/x/net/trace

Tracing with zipkin:

docker run -d -p 9411:9411 openzipkin/zipkin

Cert generation

openssl req -newkey rsa:2048 \ -new -nodes -x509 \ -days 3650 \ -out cert.pem \ -keyout key.pem \ -subj "/C=US/ST=California/L=Mountain View/O=Your Organization/OU=Your Unit/CN=localhost"

Running the App

run server

go run server.go -tls -cert_file ../certs/cert.pem -key_file ../certs/key.pem

run client

go run main.go -tls -ca_file ../certs/cert.pem

Docker Builds

Buidl container

docker build -t obiwan007/usersrv . docker push obiwan007/usersrv

Run service

docker run -it --rm -p 10000:10000 --name usersrv obiwan007/usersrv

docker run -it --rm -p 8080:80 custom-nginx

Clustering with Kubernetis

Kubectl

https://kubernetes.io/docs/tasks/tools/install-kubectl/

kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.0.0-beta8/aio/deploy/recommended.yaml kubectl proxy

https://kubernetes.io/docs/tasks/access-application-cluster/configure-access-multiple-clusters/

Local Development

Minicube Installation

https://minikube.sigs.k8s.io/docs/start/

brew install minikube

Show Dashboard

minikube dashboard

Deploy to minikube

kubectl create -f userservice.yaml

Usefull commands

k kubectl apply -f config/nginx.yaml

k get pods

kubectl expose deployment nginx-deployment --port=80 --type=LoadBalancer

Rancher fluentd secret config for logging; <parse> @type regexp expression /^(?<time>.+) (?<stream>stdout|stderr)( (?<logtag>.))? (?<log>.*)$/ time_format %Y-%m-%dT%H:%M:%S.%N%z </parse>

Database

Install

apt-get install postgresql

Change permissions:

sudo vi /etc/postgresql/12/main/pg_hba.conf

host all all 0.0.0.0/0 md5 host all all ::/0 md5

user postgress

sudo -u postgres psql

postgres=# \password

create databases

sudo -u postgres createdb <dbname>

clientsrv projectsrv timersrv usersrv

https://mx.ogseu0itmevmpymz.myfritz.net:443/auth/callback

obiwan007 / usersrv