Let’s Talk About Time - Data Driven Organic Monitoring
summary
An internal platform called Slant ingests immutable event streams into Elasticsearch via an async pipeline and exposes a REST API; it models incidents as a finite state machine and derives materialized views so timelines, virality scores and arbitrary Lucene queries drive grouping and escalation. The system claims it automatically merges related alerts, surfaces service impact and support tickets, and identifies root cause in 80–90% of cases, reducing alert noise and enabling contextual drilldown.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 0:15Consensus Slide
- 2:30What is an Event
- 4:35Why Events Matter
- 5:57Incident Center of the Universe
- 7:48Continuous Time
- 9:09Events at Scale
- 10:05Outputs
- 10:18Aggregation
- 12:44Other Advantages
- 16:29Adding Mathematical Functions
- 18:12Statistics
- 18:44Breach offender list
related talks
43mObserving from Incidents
41mCustomer Centric Observability at Scale Leveraging AIOps & OpenTelemetry
43mBeyond Logs: Real-Time Anomaly Detection with Machine Learning
29mData Observability: From 1788 To 2032
19mLeveraging Analytics for Technical Efficiency and Enhanced User Experience
01h 00mElasticsearch: you know, for more than search
42mMachine Learning In Observability Systems: From Insights To Action
12mA Principled Approach to Monitoring Streaming Data Infrastructure at Scale