I have microservice stack working locally - Docker, Eureka, ConfigServer, Spring Boot and multiple PostgreSQL database. However, now that it is time to deploy to my AWS account I am unable to find some good documentation on how to do this via CDK V2. I really want the ability to deploy this way as I will have multiple duplicate (DEV/QA/PROD) environments.
I am struggling on how to build this via CDK - multiple subnets (one for db and one for services), Route53, S3, IAM etc. Most examples I find are for a single service, but not how to create the RDS and connect them.
Can anyone point me to some good tutorials or examples so I can move forward?
I have some questions about terminology that I should use. I have searched up the components that I am using, but I am still not sure. Can you please check if these are right way to describe each component? If not, can you please revise it?
I develop microservices using Spring Framework (Spring boot).
I deploy components on cloud via Kubernetes and Docker. (I understand Docker is used for containerizing and Kubernetes is for orchestrating multiple containers. Is it right to say "I deploy on Kubernetes or via Kubernetes"?)
CI/CD pipeline using Jenkins and Azure DevOps. (Project uses Azure DevOps for triggering CI/CD in Jenkins)
Please note that this project was already there when I joined the team, and I am new to all of these concepts. I understand what they do briefly and I know what each terminology means on entry level, but I just want to make sure if I am saying these in right ways.
I would say that you deploy services, not components, but your team might have its own terminology.
You do deploy on Kubernetes.
Docker is used to create and manage containers and container images. Kubernetes does not use Docker but can use images created via Docker to deploy containers in Pods (via the OCI format)
The rest seems right to me :)
In Kubernetes an app as a unit is called Service. It has components reachable under it, for example: deployment, pod, etc. This set of components could be for microservice or service, but terminology just a service.
You can list all services using kubectl get services
The Docker is related here only because pod has a container or sometimes few of them. So Kubernetes supports Docker and its images. It's most popular container runtime in 2021, but not the only one for Kubernetes.
To combine everything, the right sentence to me would be:
I deploy a service on Kubernetes
I deploy a service on Amazon EKS
I deploy a service on Azure Kubernetes
etc.
Adding "cloud" doesn't add so much value as Kubernetes platform is same regardless of cloud provider chosen. Of course, there are some differences in cloud, but not in Kubernetes core.
Docker doesn't fit here and usually discussed when going into details, especially about images, building them, etc.
Also off topic, but even pod is more used than container, because in Kubernetes you access the pod (while in Docker Compose, you access the container). Of course, you can access Docker container if there are multiple of them setup in a pod and container engine is Docker in this context.
I develop microservices using Spring Framework (Spring boot).
To understand in the most basic form - Microservices is an architectural pattern for developing applications, so there can be many ways to implement this pattern (which is nothing but developing applications) and Spring Boot is one of them. More precisely you should say "I develop microservices using Spring Boot"
I deploy components on cloud via Kubernetes and Docker. (I understand
Docker is used for containerizing and Kubernetes is for orchestrating
multiple containers. Is it right to say "I deploy on Kubernetes or via
Kubernetes"?)
Precisely speaking you are not correct here - you understand the purpose of Docker and Kubernetes but putting it wrongly w.r.t. application and its deployment.
Do read this, where it is clearly mentioned that K8s doesn't build or deploy code. When the official documentation is clearly saying that K8s doesn't deploy code, then how you can say that I deploy via K8s?
I would say you shouldn't put it using either of the statement you mentioned, rather you should say - "My application runs (or is deployed) in Cloud on a Kubernetes cluster (or GKE or EKS or AKS)".
CI/CD pipeline using Jenkins and Azure DevOps. (Project uses Azure
DevOps for triggering CI/CD in Jenkins)
Looks fine to me.
Please note that in general you can say anything as long as the other person is able to understand, so in general all your statements might sell well. But I guess you were looking for "precise" statements, and I gave you my thoughts on the same.
Guys here are right but I would like to add that when speaking of Kubernetes components you usually think of Control Plane Components or Node Components and when speaking of smaller units the Services and Pods comes to mind first. Note that a Kubernetes service is an abstract way to expose an application running on a set of Pods as a network service. So addressing your main question:
Is it correct to say I am deploying components via Kubernetes?
It would be better to say that you are deploying a service/application (running) on Kubernetes (on AKS/AWS/GCP/etc).
I am migrating a Spring Boot application from PCF to AWS ECS which is currently using cloud config server reading properties from git repo, and aws RDS. Now in ECS is there a way we can implement config servers along with refresh in ECS AWS.
I think ECS operates on a different level.
Spring Cloud Config server is a solution that works especially good with spring boot based applications. For example, a refresh option that you've mentioned is implemented as a special Scope which is purely a spring (applicative) thing.
On the other hand, AWS ECS (stands for Elastic Container Service) provides a way to work with containers in a general sense (with scaling and everything). It doesn't require the containers to be spring-based or even java based.
So, I think you might want to consider keeping a spring boot driven microservice for config server just like you have now, but wrap it into docker container and deploy it in AWS ECS
We are currently working on a approach which will work with collaboration of multiple microservices to provide us with solution.
Dataflow example:
These services will be deployed in ECS with ALB in front of it. I want to stitch together a workflow for the system and needed help with the following:
We plan to use Route 53 DNS service for the Service Discovery. I know that conductor uses Eureka client for service discovery. Is it possible just to swap out of use DNS based implementation rather than Eureka??
Can Conductor be deployed as docker images in an ECS cluster? Can I create this internal to ECS cluster and not deploy it on EC2 servers? Any documentation that will help? Do we maintain Docker Images or we need to create our own?
I am trying to build a Apache Ignite Cluster on Pivotal Cloud Foundry environment as follows.
Created a Spring-Boot app that starts a new Ignite node, Deployed it on on Cloud Foundry ( ex. Ignite-Node1)
Created another Spring-Boot app which will also starts a new Ignite node, Deployed it on on Cloud Foundry ( ex. Ignite-Node2)
Now, even though both the apps are running in a same CloudFoundry Env, they are not forming the Ignite Cluster as they are not able to discover each other.
Apache Ignite documentation provides examle configuration details for AWS, Google cloud , however there are no examples for Pivotal Cloud Foundry.
Can somebody provide me with an example configuration on how to get the Ignite Cluster running on Cloud Foundry environment.
Srini
Container-to-container (app-to-app) networking is not supported on Pivotal Cloud Foundry, although it is possible to enable.
By default all communications must be made 'via the front door' through Cloud Foundry's router, either via HTTP or on a TCP port. One cannot choose which instance of an application to hit. This is due to change with the Container Networking initiative, the progress of which you can check on Pivotal Tracker. There is a detailed design document which is publicly available.
An alternative approach which is more appropriate for data services is to deploy them using BOSH. As a PCF user, you may wish to consider creating a PCF Tile for Apache Ignite.
My company has been helping Hazelcast create a PCF Tile that will create dedicated-VM clusters on-demand. Perhaps you could consider making use of Hazelcast instead?