Data Observability - System Health to Business Health
summary
Claims measurable metrics: Data Downtime as time series of partial/error/missing/inaccurate data, Data Flowtime as source-to-ready latency, and a contractual 99% data SLA tied to those measures. Architecture uses a shared-nothing data plane plus shared control plane and a programmatic governance runtime that injects quality rules into tenant-specific pipelines. Platform runs thousands of pipelines (~4,000 realtime, ~10,000 extraction jobs) with ~5% monthly growth and reports 99.95% infra uptime; several live incidents show the approach detects, isolates, and fixes problems within minutes.
AI-written summary · about
chapters (14)
- 0:00Data observability intro
- 1:30Data companies evolution
- 7:09Zotap’s CDP challenges
- 11:27Data products vs platforms
- 13:51Customer expectations in data
- 16:00Observability architecture at Zotap
- 19:29Data downtime explained
- 25:07Flow time measurement
- 27:55SLAs for data reliability
- 31:43Governance with analytics
- 33:31Observability in action
- 37:45Future tech directions
- 39:10Trust from observability
- 40:12Closing & Q&A
related talks
29mData Observability: From 1788 To 2032
47mObservability - Holistic Quality across Software Systems
12mMastering Observability in Distributed Systems
35mApplying Data Quality Monitoring
26mDataOps: Accelerating Digital Transformation
19mObservability for Data Pipelines: Monitoring, Alerting & Tracing Lineage
57mWhat Is Data Observability and What Benefits Does It Bring to Data Teams?
26mMeasuring Success - Observability and More!