Customer Centric Observability: How Intuit Reduced Time To Detect Customer Impact From 30+ Minutes To Under 3 Minutes.
Talkat KubeCon + CloudNativeCon North America 2022 · on Thursday, 2022-10-27· at 17:25· Portside Ballroom (Room 260)·
summary
Claim: MTTD falls from over 30 minutes to under 3 minutes by detecting Failed Customer Interactions (FCI) end to end. Implementation uses a tiny front-end library that tags important interactions, generates OpenTelemetry trace context, and sends spans to an OpenTelemetry collector, Grafana Tempo for traces, Kafka + Beam/Flink for stream processing, Apache Druid + Apache DataSketches (HyperLogLog) for cardinality and OLAP, Wavefront for alerts, and Numa for anomaly score detection. Demo shows exemplar traces isolating a DB span, 30 billion interactions and ~$2M vendor savings.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:23Al Driven Expert Platform
- 1:16Open Source & Open Collaboration
- 3:35Experience Availability
- 4:50Failed Customer Interactions (FCI) Platform
- 6:23FCI Instrumentation
- 8:04High Level Design
- 10:20Quantify Impact
- 11:35Quality Signals
- 20:32What's next
- 21:47Customer Centric Observability
related talks
10m⚡ Lightning Talk: Unlocking Customer-Centric Observability: A Case Study of OpenTelemetry To Reduce MTTD < 3 Mins
41mCustomer Centric Observability at Scale Leveraging AIOps & OpenTelemetry
19mOops my app performance sucks... what if observability could be a game changer?
28mScaling Observability for Real-Time Personalization Engines
50mIntuit Journey To Unified Observability at Scale: Challenges, Be...
41mResilient Observability at the Retail Edge: A Lightweight, Scalable, and...
30mPanel: Telemetry That Matters
12mMastering Observability in Distributed Systems