Ultimate Guide to Microservice Architecture on Kubernetes
Download as PPTX, PDF2 likes784 views
This document provides an overview of microservice architecture on Kubernetes. It discusses: 1. Benefits of microservice architecture like independent deployability and scalability compared to monolithic applications. 2. Best practices for microservices including RESTful design, distributed configuration, client code generation, and API gateways. 3. Tools for microservices on Kubernetes including Prometheus for monitoring, Elasticsearch (ELK) stack for logging, service meshes, and event sourcing with CQRS.
Ultimate Guide to Microservice Architecture on Kubernetes
- 1. Ultimate Guide to Microservice Architecture on Kubernetes Huseyin BABAL Software & DevOps Consultant
- 2. Who Am I? Software & DevOps Consultant Ex-Sony and Ex-eBay Engineer (Microservice Transformation Project Architect) Google Developer Expert on Web Technologies Organizer of Docker Istanbul, NodeSchool Istanbul, DevOps Underground meetups
- 3. Who we are ? ● Microservice Transition ● DevOps as A Service ● Test Automation as A Service ● Managed Kubernetes
- 4. Who we are ?
- 5. Customers
- 6. Ready to Start ?
- 7. Once upon a time while we are in monolithic app days
- 8. After switching to Microservice Architecture ...
- 9. And yes, the truth is, only the name Microservice Architecture cannot solve your architectural problems. You need to consider applying best practices to Microservices to do it in an efficient way
- 10. #1 Try to Reach Glory of REST
- 11. Leonard Richardson’s Maturity Model
- 12. #2 Use HATEOAS
- 13. HATEOAS Hypermedia As The Engine Of Application State
- 14. User Model
- 15. Article Model
- 16. Article Resource
- 17. User Resource
- 19. User Controller
- 20. The Result
- 21. Pro Tip If you don’t want to convert models/dtos to resources manually, you can use Resource Assembler.
- 23. Why Not Project Specific Configs? ● Sensitive data walks around Git ● Unable to inherit common properties like spring.main.banner-mode=OFF
- 24. How to Centralized Config? ● Consul can be used to keep config data as Key/Value ● Create a project for just keeping project configurations. ● Git2Consul for sync configuration to Consul
- 25. Architecture Git2Consull Daemon Config Project Git push new config change Polling Sync configs to Consul User
- 27. Spring Boot Config (Test)
- 28. Git2Consul npm install -g git2consul Create a file called git2consul.json and add necessary config git2consul --endpoint <consul_host> --port 8500 --config-file git2consul.json
- 30. Pro Tip By default, Spring Boot refreshes its context on config change on Consul. This may cause down time problems, so disable config change watching with following.
- 31. #4 Client Code Generation
- 32. How? You can either use Swagger to generate your client code on any supported language, or feign client with a little annotation and client side load balancing with Ribbon.
- 33. Feign Client
- 34. Swagger Doc
- 35. Swagger Doc Now you are able to access; http://your_api/swagger-ui.html for api documentation http://your_api/v2/api-docs for json specification of API doc.
- 36. Swagger Codegen
- 37. Pro Tip The best place to generate api client is while Jenkins build section. ● If you are deploying a feature to non-prod environment you can generate client library with snapshot version and push to nexus. ● If you are deploying a feature to prod environment, you can generate client with stable and push to nexus artifactory
- 38. #5 k8s Warm-up
- 40. deployment.yml
- 41. service.yml
- 42. Kubectl configuration Kubectl is a client app for k8s api server in order to manage k8s cluster. If you use minikube, your kubectl will be automatically configured, and it is different for other cloud providers.
- 43. Simple deployment git clone <http://project> cd <project> kubectl apply -f k8s
- 45. #6 CI / CD
- 50. Deployment Types ● Rolling Update ● Canary Deployment ● Blue / Green Deployment
- 51. Rolling Update Deployment resource on k8s uses RollingUpdate strategy by default. Within this strategy, pods deployed one by one instead of taking entire service down.
- 52. Canary Deployment You deploy an experimental feature and allow small amount of request traffic to this deployment. You increment the size of traffic and after a while, canary replaces the production one
- 55. After a while... You confirmed that, the feature on canary deployment works, replace prod image with canary one and delete canary deployment
- 57. Blue & Green Deployment In this strategy, there will be 2 environments with same properties except application version. The current version will be called blue and new version will be green. Just update ingress rules to redirect traffic to green deployment.
- 58. #7 Monitoring
- 60. Prometheus Prometheus is capable of collecting metrics from known sources like cAdvisor. Prometheus is mainly used for collecting metrics and alert manager to notify you on any kind of problem
- 61. Prometheus Operator Hopefully, CoreOS team developed a project called prometheus operator to collect k8s specific metrics automatically. https://github.com/coreos/prometheus-operator
- 62. Getting Started You can install Prometheus with Helm charts
- 63. Visualization Prometheus lets us to keep track of external services by using some endpoint via exporters. To visualize metrics, we will use grafana.
- 65. Alert Manager
- 66. #8 Logging
- 68. Logging Types ● Node Level Logging ● Cluster Level Logging
- 71. You can use several technology to send your logs to logging backend. It can be Graylog, ELK, etc...
- 72. logz.io
- 73. Kubernetes Setup You can run a daemonset to send your logs to logz.io as stated here: https://github.com/DanielBerman/k8s_logging_files/blob/master/logging_logz/daemons et-logz.yaml
- 74. #9 APM & Service Mesh
- 75. Why to Monitor Service Metrics? Beautiful graphs and dashboard fetched from log resources may not be helpful for you every time when you face a difficult issue. You may need to see your service insights to find the root cause.
- 76. Tools Can Be Used NewRelic, AppDynamics, DynaTrace, Zipkin can be suggestion for your APM monitoring. However, in a containerized microservices world, you may need to have a tool that works in real-time and has some AI capabilities.
- 77. Instana Instana is an AI Powered Application and Infrastructure Monitoring
- 84. #10 API GATEWAY
- 85. Microservice Patterns? In most cases, we have foundation services to access data, and aggregation layer to aggregate data to serve specific clients. According to my experience, you can handle microservice division in two ways. (There are many more, but those two are general)
- 86. Aggregation Layer Service 1 Service 2 Service 3 DB 1 DB 2 DB 3 Cache 1 Cache 1 Cache 1 Aggregation Service Load Balancer
- 87. Aggregation Layer When client wants to get data that contains data from multiple services, it consults to endpoint on aggregation layer. GET /users/1 User Response = Data(Service1) + Data(Service2) + Data(Service3)
- 88. Proxy Pattern In this pattern, you can use API Gateway to proxy requests to downstream microservices. Kong, and Tyk are the more popular options for API Gateway
- 89. Service 2 Service 1 Service 3 Service 5 Service 6 Service 4 Service 7 Load Balancer
- 90. ● Proxy ~3000 req/sec with 65 ms latency ● Proxy ~2000 rew/sec with 85ms latency including key validation, security check, quota management ● 2-Core 2GB Virtual Server
- 98. #11 Event Sourcing & CQRS
- 99. CRUD Nature In services, we always try to change a state of a resource at one time in one transaction. When more instance needed, we started to use distributed transactions by applying 2- Phase-Commit protocol in Extended Architecture (a.k.a XA). Whenever you do a change current state of an entity, old transaction logs gone.
- 100. Event Sourcing Instead of saving current state of entity, save state-changing events belongs to entity in a time series format You can replay events within event store, even you can reproduce a bug if exists on your production! There is no update to past state of an entity, instead of use an updated event entity in new row.
- 101. Product Events Product(‘Computer’, 1000 USD) Product(‘Laptop’, 1000 USD) Product(‘Laptop’, 900 USD) ProductMetadataUpdateEvent ProductPriceUpdateEvent
- 102. How? You can use Apache Kafka, Cassandra, etc … to keep you revents to make them available to any kind of consumer. Just save event on any action for example user creation, but you don’t have to return response immediately. Event handlers, handle event to populate their views to make available to query
- 103. CQRS Command and Query Responsibility Segregation A Command alters the state of an object, but does not returns data A Query returns data, but does not alter state of object
- 104. Product Service Create Product Update Product Delete Product ProductCreateHandler Topic per Event Or Topic per Object and event type in log Apache Kafka Database APIGATEWAY Create Product Request Read Product Request
- 105. Frameworks