I found that the bottleneck of zipkin is collector and API, are these two components stateless so i can deploy multi collectors and multi API?
I want to deploy zipkin in kubernetes.
Yes, they are both stateless. You can deploy them using whatever horizontal-scalability construct is available to you.
Related
We are evaluating Dapr for our microservice framework. We will be using Kubernetes. One of the advertised selling points for Dapr is service invocation and service discovery. Doesn't K8s already offer service discovery out of the box?
Short answer: Yes (Kubernetes offers Service Discovery)
While there may be several patterns (and tools being the implementation of these patterns) for Service Discovery, Kubernetes offers service discovery at its core through the Service objects avoiding any needs of using a particular technology or tool to achieve a basic Container-Managed runtime environment.
You can read more on Kubernetes Services in the official documentation.
It is worth noting that dapr is a platform agnostic portable runtime that does not depend on Kubernetes and its core Service Discovery feature.
It offers more features than simply discovering your services (it has been usually compared to Service Meshes tools as the tend to look out being the same):
It offers transparent and secured service-to-service calls
It allows Publish-Subscribe communication style
It offers a way to register triggers and resource bindings (allowing for function-as-a-service development style)
It offers observability out-of-the-box
...
I want to run a restapi server, a spring boot app, to kick start other docker containers inside of the same kubernetes cluster. I found these 2 packages:Spring Cloud Kubernetes and Kubernetes API Client. I am confused on what the difference between the two and which one should I use.
With Spring cloud cal version (2020.x.x), spring cloud Kubernetes comes with two implementations - Fabric8 and Kubernetes Java Client. See here.
NOTE - Prior to spring-cloud cal version spring-cloud-kubernetes used fabric8 impl
Both implementations works fine. While there are many, here is one of the blog comparing the two. In general Fabric8 seems to have more features.
I am not sure if the Spring community prefers any particular impl. Both projects are active.
Spring example here.
I too have an example using Frabric8 on github.
We currently deployed our Spring Boot Application in GKE(Google Kubernetes Engine) and we are currently using cloud endpoint to secure our web services. We have 11 web service developed which will be consumed by external clients. Is there any way i check the SLO (times, performance) of a webservice in cloud endpoint or in stackdriver.
You might want to check:
Spring sleuth
Jaeger operator
Jaeger is a opentracing standard and can help understand the values, and sleuth is a tool to integrate with spring, there are several options, you might want also to consider opencensus
First you need to expose metrics from your applications. Spring Sleuth is a great choice if you're using Spring Boot.
Then you need to collect the metrics and visualize them. Google provides a tool for that called Stackdriver Trace. It can also do metric-based alerts. You can find a sample setup for your use case here.
There are other performance monitoring services such as Dynatrace or Datadog.
If you want a self-hosted solution, you can use Zipkin which is inspired by an internal Google system called Dapper.
Have you looked at Google cloud console UI? Its "Endpoints" tag should show all services your project is running.
Need some suggestions & tips, Please let me know if anyone build the infrastructure on K8s with Oracle Weblogic, DB & SOA (OSB, BIPEL, B2B, BAM).
If anyone build the infrastructure please let me know the benefits & drawbacks. Is it recommended to build it.
Thanks
If you follow best practices; do not install these on Kubernetes. Kubernetes is very good at for micro-services and stateless applications. Of course Kubernetes support stateful applications like cassandra / nodejs ..etc applications you have a chance to deploy applications on Kubernetes using StatefullSets. Kubernetes treats statefull sets different than other resource objects. A StetfullSet Pod is always scheduled on the same machine as long as you do not delete them.Please take a look at the link https://kubernetes.io/docs/tutorials/stateful-application/cassandra/ You can deploy stateful applications on Kubernetes using the link above
https://blogs.oracle.com/weblogicserver/weblogic-on-kubernetes,-try-it
https://blogs.oracle.com/weblogicserver/automatic-scaling-of-weblogic-clusters-on-kubernetes-v2
We ran weblogic applications on docker swarm but not on kubernetes. All the configuration on kubernetes for stateful applications is pretty easy compared to docker swarm.
We are planning to integrate a microservices platform which is running in Kubernetes and the architecture will be similar to the below reference architecture.
https://www.ibm.com/devops/method/content/architecture/microservices/1_0
What are the core components which can be replaces by Istio on the above Kubernetes microservices architecture?
I understand it can help in request routing, traffic management, policies and control.
If I need to migrate one of the similar project what are the components which is available in Istio which can be replaced in Kubernetes or vice versa?
It doesn’t replace any kubernetes core components. ISTIO is add-on for kubernetes to manage the micro service.
In the IBM article you can use ISTO for routing instated of Ingress Controller.