Observability & Monitoring in Cloud Native Telco World – Tying It Together
Talkat Cloud Native Telco Day Europe 2023 · on Tuesday, 2023-04-18· at 10:50· Hall 7, Room A | Ground Floor | Europe Complex·
summary
Integration of a cloud‑native 5G core into Deutsche Telekom’s existing monitoring stack is underway and partially working: metrics run on Prometheus/Thanos/Grafana per cluster but face 70,000 time series causing Thanos queries to spike to ~4.6 GB RAM and 20+ second responses; logs move through rotating agents (Fluent Bit → Vector → Filebeat) because of security, support and multi‑endpoint limits; networking uses Istio with JSON access logs plus optional Zipkin/Jaeger tracing; centralized sinks and ML detect anomalies but alert unification, cardinality and automation remain unresolved.
AI-written summary · about
chapters (19)
- 0:00Observability in cloud native telco
- 0:43Agenda and ecosystem
- 1:49Events logs metrics alerts traces
- 2:11Log shipping in Kubernetes
- 3:55Security in log shippers
- 5:19Prometheus cardinality issues
- 6:48Alerts across multi-vendor systems
- 7:50Git commits as signals
- 8:11Network tracing
- 11:28Data troubleshooting steps
- 12:51Anomaly detection with ML
- 13:57Evaluating KPIs
- 15:00Automating legacy interactions
- 16:24Chaos testing resilience
- 17:30Deployment strategies
- 18:55Real-world deployment challenges
- 19:16Tooling and future enhancements
- 20:00Key takeaways
- 21:04Q&A
related talks
27mObservability Brings Clarity in 5G World
34mMulti Cloud Observability for 5G
24mObservability at Scale: How Open Systems Collects Telemetry from Over 10'000 Edge Devices Worldwide
25mOpen source observability for private cloud: mission impossible or not?
55mCloud-Native Observability with OpenTelemetry
43mDistributed Monitoring: How to Understand the Chaos
22mHarnessing Observability for 5G Performance: eBPF and OpenTelemetry I...
42mDistributed Monitoring: How to understand the Chaos