The Invisible Tax: How Data Format Conversions Drive up Telemetry... Cijo Thomas & Joshua MacDonald
Opentelemetry Telemetry pipelines Observability Data format conversion Cloud native Pipeline optimization Collector performance Cost optimization Marshaling overhead Kubecon
This talk presents quantitative research on the hidden costs of data format conversions in telemetry pipelines. Cijo Thomas and Joshua MacDonald from Microsoft measure compute spent on pure format conversion versus value-generating processing in OpenTelemetry SDK and Collector pipelines, showing how these costs scale at terabyte scale. The session is for observability engineers and platform developers looking to optimize their telemetry pipeline costs.
Related videos
The Speed of Metrics, the Fidelity of Traces: Architecting Post-Collection Aggregation - Zack Owens
Grafana Tempo: The distributed tracing journey to 3.0 (June 2026 Community Call)
The Full Picture: Visualizing Service "Fullness" To Rethink Saturation Prevention - Tal Nordan
Keynote: 10 Million Spans Per Second: Lessons From Scaling OpenTelemetry at Reddit - Trevor Riles
OpenTelemetry GenAI in Practice: What the Spec Says Vs. What You Actually See - Zach Groves, Datadog
Panel: Telemetry That Matters - Diana Todea, Antonio Jimenez Martinez & Laura Luttmer
Taming Tenancy, Cost and Architecture at Collibra Through OpenTelemetry and Our... Alex Van Boxel
Beyond Dashboards: Cultivating the Observability Mindset - O. Hope Amaechi-Okorie, JSON Schema
DNS Tracing & Metrics Via eBPF in OpenTelemetry- Endre Sara, Causely & Nikola Grcevski, Grafana Labs
Rab Ne Bana Di Jodi - Why Platforms & AI Need Each Other - Ram Iyengar, Cloud Foundry Foundation
20 minutes to 2 minutes: How Home Depot automated incident responses with OpenTelemetry
