Stability Patterns for Microservices
- 1. Stability Patterns for Microservices Petru Flueras Software Architect - Visma @pflueras Codecamp Timisoara, May 2019
- 3. Microservices are usually processes that communicate over a network to fulfill a goal using technology-agnostic protocols such as HTTP
- 4. Fallacies of distributed computing 1. The network is reliable 2. Latency is zero 3. Bandwidth is infinite 4. The network is secure 5. Topology doesn't change 6. There is one administrator 7. Transport cost is zero 8. The network is homogeneous source: https://en.wikipedia.org/wiki/Fallacies_of_distributed_computing
- 5. Can we build stable services over unreliable networking?
- 6. Unstable Services ● Scarce resources ○ Memory ○ CPU ○ Thread pools ○ Sockets ○ Connection pools ● External conditions ○ Integration points ○ Downstream network calls
- 7. In addition to unreliable network there are turbulent conditions at service level
- 8. Pitfall #1
- 9. Pitfall #1 Cascading failures, Blocked threads, Slow responses
- 10. #1 Timeouts Protect your system from someone else’s failure
- 11. #2 Retries (Idempotent!) ● Fail over temporary failures ● Retry with exponential backoff policy ● Idempotence: is the property of certain operations that can be applied multiple times without changing the result beyond the initial application ● Idempotence = safe to retry!
- 12. #3 Circuit Breaker Gives the opportunity of external service to recover source: https://www.ebayinc.com/assets/Uploads/Blog/2015/08/circuit_breaker_state_diagram.gif
- 13. Pitfall #2
- 14. #4 Bulkhead Isolate elements of an application into pools so that if one fails, the others will continue to function
- 15. #5 Throttling ● Over capacity protection? ● μService is overwhelmed with requests ● Avoid slow responses ● You can’t out-scale the world ● Control the consumption of resources used by a service ● Control maximum number of concurrent requests ● A queue between accepting connections and processing requests ● Can be used for ‘auto scaling’
- 16. Testing Stability ● Chaos Monkey (Netflix) ○ resiliency tool that helps applications tolerate random instance failures ● Simian Army (Netflix) ○ a suite of tools developed by Netflix to test the reliability, security, or resiliency ● Google DiRT (Disaster Recovery Testing) ● Chaos Engineering ● Monitoring!!!
- 18. Thank you!