Simple GRPC based Microservice Server & Client
==============================================
This project contains end to end functionality to test simple Go grpc based microservice depoloyed in K8s.
It has 3 components
i - cli to access the compute service (add, mulitpy)
ii - server cod & docker Manifest
iii - kubernetes deployement files
Steps To deploy the server in the K8s cluster
----------------------------------------------
***refer to pre-requesties section before running the k8s deployement
- To deploy the server, service object
kubectl apply -f grpcTest/K8s/grpcServerDeploy.yaml
- To deploy the ingress
kubectl apply -f grpcTest/K8s/voyager-ingress.yaml
Steps To run the CLI Client
----------------------------
go get github.com/tamilhce/grpcTest/cliClient
cd <GOPATH>/src/github.com/tamilhce/grpcTest/cliClient
go build -o <GOPATH>/bin/cliClient .
<GOPATH>/bin/cliClient --address <serviceIp> --port <servicePort> --ops <operation> int1 int2
i.e /home/aviuser/go/bin/cliClient --address 10.79.58.24 --port 80 --ops sum 10 20
Steps To Build New Image (optional)
========================================
**proto**
cd github.com/tamilhce/
protoc --proto_path=proto --go_out=plugins=grpc:proto service.proto
**server docker image**
cd github.com/tamilhce/grpcTest/server
docker build -t <docker-id>/grpctest
docker docker push <docker-id>/grpctest
**client binary**
cd github.com/tamilhce/grpcTest/cliClient
go build -o pkg/cliClient
To Run the Unit Test
====================
cd github.com/tamilhce/grpcTest/server
go test -v
Pre-requisites
===============
i) voyager Ingress controller shoud be installed in prior
https://appscode.com/products/voyager/v12.0.0-rc.1/setup/install/
Reason for choosing voyager
grpc support for Nginx Ingress is available in Nginx controller version >= 0.30
In my current setup Nginx controller version is 0.25, facing some permission issue to upgrade to >=0.30
voyager - HA proxy based, simple to use & stable
Limitations
=============
ingress grpc annotations are Ingress controller specific.
In our case the Ingress controller & Service Type Load balancer should be made avilable in the cluster
Future work/ Enhancements
=========================
i Nginx Ingress controller with TLS terimination
ii Helm package for the K8s deployment
ii REST API based client implementation approach
FAQ
===
i) How to scale the application ?
By increasing the replication count in the deployment, the application can be easily scaled
ii) How to upgrade/rollback the app ?
By changing the image/config in the deployment (default strategy: Rolling update) we can upgrade or rollback to the previous version
iii) Is the application is stateful/stateless
The applciation is stateless & so it is easy to scale as a microservice
iv)is Application is TDD aware?
yes, The application pacakge is bundled with go unit test for each functions , The test cases can be run autamatically for new changes, we can extend additional end to end functional cases as part of CI/CD ?
v) Is it CI/CD aware ?
yes, The build procedure mentioned above can be added in the jenkins pipeline to create new image to be pushed into the repo,when ever a set of commit happens
vii) How to extend the eventstore to this application ?
yes, there multiple design patterns available, one common approach is adding side-car to stream a logs(std out/erro) logs to external logging system ( e.g elasticsearch)
viii) How the dependencies are managed?
By packing the applcation as container solves the dependencies & environment issues(cross-platform)
Having said this, The application is cloud-native, scalable microservice application :)