Google Developer Group Cloud - Modernizing Government Tech: Observability and Data-Driven Practices

Transcript

1. Modernizing Government Tech: Observability and Data-Driven Practices Bandung
2. Jabar Digital Service Yoga Hanggara Head of Engineering
3. Intro
6. YES, BUT Dimana bisa menemukannya? Bagaimana cara mengambilnya? Bagaimana menyimpannya? Data is the new oil. Image by macrovector_official on Freepik Image by Freepik
7. Mengambil keputusan karena suka vs tidak suka, atau HIPPO (Highest Paid Person's Opinion). Good data, good decision. YES, BUT Image by Racool_studio on Freepik Image by pressfoto on Freepik
8. Quiz
9. Apa yang membuat Anda malas menggunakan aplikasi yang dikembangkan oleh Pemerintah? 55 Transaksi/proses sama ribet dan lama 32 Data tidak aman 18 Lambat atau sering down 10 UX tidak nyaman 5 Tidak terintegrasi
10. 1.000.000 Tidak ada yang pakai aplikasi kami? Jumlah users di database
11. 64 CPU 256 GB RAM Kapasitas server kami besar Spesifikasi Server yang Dibeli
12. 0 Aplikasi tidak ada error Jumlah komplain/aduan dari user
13. Users active 500.000 Week 2 Week 3 0 Week 4 Expected Week 1 Actual
14. Server Capacity vs Loads 100% Week 2 Week 3 0% Week 4 Expected Users Week 1 Expected Capacity Actual Workload Inefficient
15. User Complaint/Error Tidak ada monitoring = Tidak ada error
16. The Role of Observability
17. “Yuk ke Jakarta naik mobil, bensin masih separo. Cukup kayaknya. Ga usah bawa uang tambahan buat beli bensin.”
19. Angka = Bahasa Universal “Banyak”, “Cukup”, “Sedikit” 1000, 100, 10
20. Banyak user yang komplain error! Terdapat 10 user yang mengalami error, sedangkan 990 user berhasil. Kita akan mendalami informasi lebih lanjut pada 10 user tersebut. Panic-Driven-Development
21. Aplikasi lambat diakses! Terdapat 10% user yang mengalami akses lambat, dan itu terjadi pada fitur Download. Panic-Driven-Development
22. The Role of Observability: Much like a pilot relies on instruments to navigate, assess conditions, and reach the destination safely and efficiently. Image by DCStudio on Freepik
23. Data-Driven Decision Making
24. “You can’t manage what you can’t measure.” “If you can’t measure it you can’t improve it.” Peter Drucker - Modern Business Management
25. What are you working on, and why? We’re working on Feature X, because Sam said it’s important. We’re done when Sam is OK with it. We’re working on Feature X, because we feel like it. We’re done when we don’t feel like it anymore. We’re working on Feature X, because we think it’s going to give impact Y, which matters to the company because of Z. We’re done when the metrics have moved.
26. Journey
27. 1. Culture, People. 2. Measure what matters. a. Business metrics. b. Engineering metrics. 3. Observability method & tools. 4. Dashboard, alerting, forecast. a. Operational b. Analytics Journey https://www.peoplebox.ai/blog/takeaway-famous-okr-books/
28. Measure what matters Business Metrics (Conversion) Engineering Metrics 1. User Acquisition 2. User Active, Retention 3. Order Created 4. Transaction Success 5. Video/Song Played 6. Ads Clicked 7. Ticket Closed 1. Response Time 2. Error Count 3. Throughput (Request per-second) 4. Application Performance Index 5. Automation Test Coverage 6. Security Tickets 7. Defects Count
29. Observability Method & Tools Ingest, extract, transform, load (ETL)
30. Dashboard, alerting, forecast.
31. Dashboard, alerting, forecast.
32. Dashboard, alerting, forecast.
33. Pada sebuah pagi hari yang cerah, saat tim sedang sarapan “Punten apakah TTE sedang bermasalah?”
36. Today Data Assistant: AI
37. Thank you