Shrinking the Observability Bill
summary
Claims a concrete 40% cost reduction from dropping health‑check logs, switching to Prometheus histograms and moving cold data to cheap object storage. Presents a framework: start from SLOs, keep metrics > traces > logs, tune retention and aggregation, and avoid high‑cardinality labels. Details Kubernetes targets: control‑plane metrics (API latency, HC lag), node heartbeats, pod‑level filtering, and network sampling with eBPF. Recommends tools like Prometheus + Thanos, Fluent Bit, and OpenTelemetry filtering, and sets an observability budget tied to $/GB and metric series.
AI-written summary · about
chapters (9)
- 0:00Introduction and Agenda
- 1:26Understanding Observability Costs
- 3:46Framework for Cost Optimization
- 4:59Kubernetes Monitoring Deep Dive
- 6:39Adopting Open Telemetry
- 9:12Case Study and Practical Implementation
- 11:42Cost Optimization Checklist
- 12:57Effective Alerting Strategies
- 14:43Recommended Tools and Final Thoughts
related talks
32mUnderstanding Observability in Kubernetes
11mKubernetes Observability
06mObservability Diet: Your 5-Step Plan To Trim the Data Fat
41mResilient Observability at the Retail Edge: A Lightweight, Scalable, and...
30mBut Wait! There's...Still More‽ - Observability Data Volumes and Strategies for Managing Them
10mA Simple and Practical Guide to Obser...
36mMetadata: The Secret Sauce for Full Observability
35mDriving down the cost of observability