Lightning Talk: A Drop-in System To Accelerate Metrics Observability by 100x Us...
summary
Claims 100x reduction in query cost and latency by adding a drop-in proxy that implements the Prometheus HTTP API and ingests raw metrics via remote write while computing streaming precomputation of compact sketches. It answers sum count rate and quantile queries from Grafana using sketch-based approximation with tunable accuracy and automatically falls back to Prometheus for unsupported queries. Prototype is in a public repo with a quickstart demo and an open‑source release planned for Q1 2026.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:01ASAP query overview
- 2:26Sketches for faster queries
- 3:49Sketches types
- 5:14ASAP as Prometheus accelerator
- 5:57ASAP architecture
- 7:23Live demo: ASAP vs Prometheus
- 8:48ASAP modes and future
- 9:30Q&A
related talks
13mLightning Talk: A Drop-in System To Accelerate Metrics Observability by 100x Us...
09mGraduated Project Lightning Talk: Using Trace Data for Monitoring & Alerting of Appl...
02mLightning Talk: Prometheus adoption stats
04mLightning Talk: Easily Observing Operators
05mLightning Talk: Improved Prometheus filtering in Grafana
05mLightning Talk: PHP and Prometheus For better Metrics
05mLightning Talk: Machine Learning Observability with Prometheus
13mLightning Talk: Observability-First DevSecOps: Building Resilient Multi-Cloud P...