Next-Gen Observability: reduce MTTR
summary
Operates across four clouds, 20 regions, 600+ Kubernetes clusters ingesting about 60 TB logs/day (3.8 trillion docs/month, ~15 PB on S3), consolidates sources into Kafka and applies an OpenTelemetry agent with source filters and ML so noisy data is removed early—claiming 20% reduction at source and 50% pipeline savings—converts high-cost logs to metrics, down‑samples health checks, and uses an AI assistant to synthesize queries, generate pipelines and automate runbooks so dashboards load faster and MTTR drops.
AI-written summary · about
chapters (8)
- 0:00Introduction and Agenda
- 0:41Current Infrastructure Challenges
- 2:12Next-Generation Telemetry Pipeline
- 3:21Deep Dive into Telemetry Architecture
- 7:42AI Integration in Telemetry
- 9:01Unified Portal and Service Mapping
- 10:48AI Assistant for Observability
- 12:27Conclusion and Thank You
related talks
50mIntuit Journey To Unified Observability at Scale: Challenges, Be...
46mObservability best practices for hybrid environments
56mFull-stack observability and application monitoring with
41mCustomer Centric Observability at Scale Leveraging AIOps & OpenTelemetry
23mWhen Kafka Goes Cloud Native: Observability That Actually Works!
12mOpen-source observability is the way forward, and here’s why
27mIngesting 6.5 Tb of Telemetry Data Daily Through Open Telemetry Protocol and Collectors
14mObservability simplified: Empowering engineers to make better software