What We Should All Worry About When Monitoring Serverless Applications

The way we develop and deploy software has changed dramatically in the last few years. This makes the ecosystems of our applications even more complicated than ever. In this session, we will go through the evolution and the monitoring challenges, while using the right tools to keep developer velocity.

Transcript

1. Monitoring serverless applications - should you worry? Nitzan Shapira, Epsagon
2. > whoami Nitzan Shapira (@nitzanshapira) Co-Founder, CEO @ Epsagon Tel Aviv Software Engineer > 12 years
3. Monitoring/Observability - why do we need it? Track service health Troubleshoot and fix Optimize performance/cost
4. From a monolith…
5. To microservices
6. Let’s talk Serverless
7. Serverless is great! Pay per use Autoscaling Development velocity
8. The era of APIs We want managed resources Applications become Highly distributed Highly event-driven
 
 Without access to any server!
9. Back to Monitoring/Observability Track service health Troubleshoot and fix Optimize performance/cost
10. Slow down! Let’s go one by one…
11. Track system health System == Functions ?
12. Functions are important Timeout Out of memory Cold start Unique challenges to Serverless
13. Functions are important
14. Track system health System > Functions ! Functions APIs Transactions
15. Serverless != Functions theburningmonk.com
16. Troubleshoot and fix Functions are not enough Need: track asynchronous events
18. Transactions
19. Tracing Asynchronous Invocations
20. Tracing Asynchronous Invocations
21. Tracing Asynchronous Invocations
22. Distributed Tracing
23. Distributed Tracing
24. Implementing Distributed Tracing Manual instrumentation •Before/after calls •At the end of each micro service •High maintenance •High potential of errors
25. Serverless apps are very distributed •Complex systems have thousands of functions •What about the developer velocity?
26. Can it be done differently in serverless?
27. Automation can help to keep up with the development speed of serverless
28. Example
29. Troubleshoot and fix
30. Monitoring serverless Stateless Limited running time Limited memory Coldstarts
31. In Serverless Time is Money
32. How much time do you really spend? Our own code API calls Infrastructure overhead
33. Let’s have a quick look 702ms
34. A real-life example $$$$$$$$$$$$$$$$ How it started…
35. Scanning functions – the easy way Scanning CloudWatch logs using AWS Lambda – every 5 minutes, save to RDS A new Lambda is spawned for every customer’s function (async) Sounds simple and fun! Poll Spawn (async) CloudWatch
36. As time flies… CloudWatch became highly throttled ➔ requests took a very long time ➔ 5K concurrent Lambdas, for 5 minutes, every 5 minutes !!!!
37. Track service health
38. Is that all we need? Probably not
39. Observability
40. Dynamic Service Map
41. Business Flows Subscribe Transfer Payment
42. Business Flows
43. Thank you! nitzan@epsagon.com @nitzanshapira