Designing for failure - Italo Vietro, N26

Designing for failure - Italo Vietro, N26

Transcript

1. Designing for Failure @italolelis Go Days Berlin 2019
2. Think about an airplane, a boing 777 to be especific Go Days Berlin 2019
3. ELMS Go Days Berlin 2019
4. Essentials Go Days Berlin 2019
5. Resilience is a Requirement, Not a Feature 1 Liang Guo Go Days Berlin 2019
6. Dependency Isolation and Graceful Degradation Go Days Berlin 2019
7. Health-check and Load Balancing Go Days Berlin 2019
8. Go Days Berlin 2019
9. If everything is OK you get... Go Days Berlin 2019
10. Go Days Berlin 2019
11. If things are not good but your app still can work... Go Days Berlin 2019
12. Go Days Berlin 2019
13. Otherwise... Go Days Berlin 2019
14. Go Days Berlin 2019
15. Self-healing Go Days Berlin 2019
16. In kube this is as simple as defining a YAML file rule Go Days Berlin 2019
17. Go Days Berlin 2019
18. Load shedding Go Days Berlin 2019
19. Circuit Breakers Go Days Berlin 2019
20. Go Days Berlin 2019
21. Retry Logic Go Days Berlin 2019
22. Go Days Berlin 2019
23. Bulkhead Go Days Berlin 2019
24. Rate Limiters Go Days Berlin 2019
25. Go Days Berlin 2019
26. Outbox Pattern Go Days Berlin 2019
27. Outlier Server Host Detection Go Days Berlin 2019
28. Service Mesh Go Days Berlin 2019
29. SLO's and SLI's Go Days Berlin 2019
30. Monitoring Go Days Berlin 2019
31. Go Days Berlin 2019
32. Distributed Tracing Go Days Berlin 2019
33. Go Days Berlin 2019
34. Go Days Berlin 2019
35. Open Census Go Days Berlin 2019
36. Go Days Berlin 2019
37. Recap 1. Always think about your dependencies 2. Dependency Isolation and Graceful Degradation 3. Load shedding and Request Controlling 4. Observalibility is not optional Go Days Berlin 2019
38. Questions and links! 4 Example application: https://github.com/italolelis/ coffee-shop 4 Link to the slides: https://github.com/italolelis/talks Go Days Berlin 2019
39. Thank you! Go Days Berlin 2019