Introduction to Continuous Profiling

Continuous profiling provides a lightweight way for monitoring and profiling web applications in real-time, allowing performance issues to be identified and resolved.

This introduction will cover the fundamentals of Continuous Profiling and how it helps detect and preemptively address bottlenecks, interpret profiling data effectively, use optimization tools, and adopt best practices to maintain optimal application performance.

Transcript

1. Introduction to Continuous Profiling
2. • Based in Lyon, France • Former designer and educator • Loves cooking and baking • Visual art, storytelling, photography • @thomasdiluccio.bsky.social Thomas di Luccio (He/Him) DevRel @ Platform.sh, Upsun, Blackfire Who am I?
3. • Deterministic vs Probabilistic observability • Collecting information and making sense out of it • Maximizing information while minimizing overhead A talk about:
4. Observability?
5. Observability Ability to witness the real behavior of an app, detect bottlenecks, improve and control performance
6. Observability Bionic glasses for your applications
7. Deterministic observability
8. Demo ⬇ www.book.b7e.io
9. request #1 request #2 request #3 request #5 request #4 request #6 request #7 request #8 request #9
10. request #1 profile request #2 request #3 request #5 request #4 request #6 profile request #7 request #8 request #9
11. Deterministic profiling • Multi-dimensional • Observe ALL functions of the designated requests/scripts • Two samples can be compared • A sample always contains reliable pieces of information • Metrics can be derived from samples • Strong overhead
12. Going beyond profiling
13. Performance tests
16. Profiles automatically evaluate all matching assertions
17. Custom metrics
19. Synthetic monitoring
22. CI/CD pipelines
23. Assessing the consequences of upcoming changes
24. Going even further
25. Multiple levels of data collection Profile: Extended trace: Monitoring trace:
26. Blackfire Monitoring frequent collection of minimal data
27. Monitoring sampling rate • Percentage of HTTP and CLI traffic to be monitored • Monitoring is both deterministic and probabilistic
28. Extended sampling rate • Percentage of the monitored traces for which more information are collected
30. Services monitoring
32. Alerting
34. Health report
36. Automatic profiling
38. request #1 profile request #2 trace request #3 request #5 request #4 extended trace request #6 automatic profile request #7 request #8 trace request #9
39. Blackfire Continuous Observability Solution Monitoring Profiling Alerting Health report Performance tests Synthetic monitoring CI/CD Integration
40. Proactively identify existing bottlenecks and the consequences of upcoming changes
41. Why would need an extra tool?
42. improve the information vs overhead ratio
43. Correlation between data collection and overhead Profile: data++ / overhead++ Monitoring trace: data - - / overhead - -
44. Could we have the best of both worlds?
45. Intermittent data collection
46. Intermittent data collection continuous profiling: data++ / overhead- -
47. Introducing Continuous profiling
54. Demo
55. Continuous profiling • Make periodic sampling in ALL of the requests/scripts • Holistic observability • Two samples cannot be compared • A single sample contains partial information • Probabilistic observability • Accurate information is derived from a large number of samples • Minimal overhead
56. Continuous Profiling Holistic quality observability with minimal overhead
57. deterministic versus continuous profiling
58. Continuous profiling is best for: • holistic view of application • minimal overhead
59. Deterministic profiling is best for: • drilling down issues • automated performance tests
60. Deterministic observability
61. Continuous profiling (more to be added soon …)
62. Build you own observability strategy Stay in control of your applications at all times
63. Thank you! Thomas di Luccio DevRel Engineer, Platform.sh thomas.diluccio@platform.sh