Kubernetes to improve business scalability and processes (Cloud & DevOps World London 2017)
1 like334 views
The document discusses a tech company's transition from a monolithic architecture to microservices using Kubernetes to improve scalability and operational efficiency. It highlights the challenges faced during this year-long migration, such as provisioning times and environment alignment, and outlines the new architecture and automated processes implemented. Key metrics indicate significant performance improvements, with quick environment creation and efficient deployment pipelines.
![Zero Downtime (1): Graceful Shutdown
13
lifecycle:
preStop:
exec:
command: ["/stop_helper.sh"]
deployment.yaml
#!/bin/bash
wget http://localhost:8002/stop
stop_helper.sh](https://image.slidesharecdn.com/kubernetestoimprovebusinessscalabilityandprocesses-cloudanddevopsworld-170619023455/85/Kubernetes-to-improve-business-scalability-and-processes-Cloud-DevOps-World-London-2017-13-320.jpg)
![Zero Downtime (2): Graceful Startup
14
private CompletableFuture run(Stream<CompletableFuture> startupJobs)
{
return allOf(startupJobs.toArray(CompletableFuture[]::new))
.thenAccept(this::raiseReadinessUp)
.exceptionally(this::shutdown);
}
JobsExecutor.java](https://image.slidesharecdn.com/kubernetestoimprovebusinessscalabilityandprocesses-cloudanddevopsworld-170619023455/85/Kubernetes-to-improve-business-scalability-and-processes-Cloud-DevOps-World-London-2017-14-320.jpg)
![[INFO]: Kubernetes really independent from Operating System
17
● independent from OS
● different OS at runtime
● criticalities only on masters
○ disable scheduling there
KUBERNETES
DOCKER
FLANNELD
ETCD
COREOS
NODE
old
KUBERNETES
DOCKER
FLANNELD
ETCD
UBUNTU
NODE
new](https://image.slidesharecdn.com/kubernetestoimprovebusinessscalabilityandprocesses-cloudanddevopsworld-170619023455/85/Kubernetes-to-improve-business-scalability-and-processes-Cloud-DevOps-World-London-2017-17-320.jpg)
![[WARNING]: Ingress controller
18
NODE 1
NODE 2
NODE 3
● Easier DNS Management
● Customizable Proxy Server
● 3rd Party Tool
● Requires External Sync
● All Requests go through it
● Reload Risks
F5
NGINX
NGINX](https://image.slidesharecdn.com/kubernetestoimprovebusinessscalabilityandprocesses-cloudanddevopsworld-170619023455/85/Kubernetes-to-improve-business-scalability-and-processes-Cloud-DevOps-World-London-2017-18-320.jpg)
Kubernetes to improve business scalability and processes (Cloud & DevOps World London 2017)
- 1. Add Title Subtitle Kubernetes to improve business scalability and processes Michele Orsi, 14th of June 2017
- 2. The Relevant and Inspiring Travel Company 2
- 3. A Tech Company to the Core 3 ● Tech Department: 300+ people ● Applications: ~150 ● Database: ~180 database schemas and ~50TB of data ● Servers: 1400 VMs, 300 physical machines ● Locations: Chiasso, Milan, Madrid, London, Bengaluru
- 4. At the beginning developers created the monolith 4 credits https://www.flickr.com/photos/southtopia/5702790189
- 5. 5 credits https://www.pexels.com/photo/gray-pebbles-with-green-grass-51168/ "... let’s break into Microservices!"
- 6. A lot of Issues 6 ● LONG Provisioning Time ● LACK OF Alignment across Environments ● LACK OF Alignment across Applications ○ weak contribution model ● LACK OF Awareness about OPS
- 7. A year-long Endeavour 7 ● Build a New, Modern Infrastructure ● Migrate the Search (Flight/Hotel) Product there ... without: ● Impacting the Business ● Throwing away our whole Datacenter
- 8. Our Core Axioms 8 ● Same Architecture across Environments ● A Common Framework to Align Software ● Centralized Monitoring/Logging, with Alerts ● Zero Downtime Deployment ● Automation Everywhere
- 9. Kubernetes: identify where is the pain 9 APP cluster APP APP APP Monitoring Datastore Alerting Queues Logging VCS Pipeline
- 11. kubernetes: our Architecture and Choices 11 APP1-PRODUCTION deployment replica-set app1-production.prd.mykubecluster.intra secret configmap POD-3POD-2POD-1 production
- 12. Monitoring, alerting and logging 12 APP1-PRODUCTION cluster graphite app carbon Grafana 4 collectdfluentd elasticsearch Graylog
- 13. Zero Downtime (1): Graceful Shutdown 13 lifecycle: preStop: exec: command: ["/stop_helper.sh"] deployment.yaml #!/bin/bash wget http://localhost:8002/stop stop_helper.sh
- 14. Zero Downtime (2): Graceful Startup 14 private CompletableFuture run(Stream<CompletableFuture> startupJobs) { return allOf(startupJobs.toArray(CompletableFuture[]::new)) .thenAccept(this::raiseReadinessUp) .exceptionally(this::shutdown); } JobsExecutor.java
- 15. Automate Everything: pipeline(s) 15 pull jar build docker (gate) QA canary (gate) QA stable (gate) PREV canary (gate) PREV stable (gate) PROD canary (gate) PROD stable ● git push ○ Continuous Integration ○ Continuous Delivery
- 16. Automate Everything: monitoring and .. alerting 16 ● (1st) pipeline deploy ○ grafana dashboard(s) ○ .. alert(s)
- 17. [INFO]: Kubernetes really independent from Operating System 17 ● independent from OS ● different OS at runtime ● criticalities only on masters ○ disable scheduling there KUBERNETES DOCKER FLANNELD ETCD COREOS NODE old KUBERNETES DOCKER FLANNELD ETCD UBUNTU NODE new
- 18. [WARNING]: Ingress controller 18 NODE 1 NODE 2 NODE 3 ● Easier DNS Management ● Customizable Proxy Server ● 3rd Party Tool ● Requires External Sync ● All Requests go through it ● Reload Risks F5 NGINX NGINX
- 19. Give me the Numbers! 19 ● 5K req/sec in the New Cluster ● 1.5M metrics/minute flows ● 10 minutes to create a new environment ● whole pipeline runs in 16 minutes ○ 4 minutes to release 100 instances of a new version
- 20. 20 .. and the cloud?
- 21. Hybrid cloud: federated cluster 21 LB on-premise cloud NODE 1 NODE 2 NODE 3 NODE 4 NODE 1 NODE 2 NODE 3 NODE 4 NODE 5 NODE 6 NODE 7 NODE 8
- 22. R&D: bring a micro-service into Google Cloud 22 APP GKE APP APP APP Monitoring CloudSQL Alerting PubSub Logging BigTable Stackdriver