Deploying OpenTelemetry Organizationally: From Proof of Concept to In-Production at Scale
summary
Pluto TV moves from many vendor-specific telemetry streams to a unified pipeline that instruments services with a drop-in library, routes via Envoy/istio to an OpenTelemetry collector, exports in parallel to Elasticsearch and OpenSearch for real production comparison, and scales from ~20M to ~68M MAU while handling ~1 GB/s logs and ~40M RPM. Engineering adds a composite propagator for B3 plus X-Request-ID, uses tail-based sampling and load‑balanced collectors with persistent storage, and lands a single dashboard correlating logs, traces and metrics.
AI-written summary · about
chapters (10)
- 0:00Introduction and Pluto TV overview
- 2:21Challenges with multiple observability tools
- 5:31Starting the OpenTelemetry proof of concept
- 10:09Building the OpenTelemetry architecture
- 14:01Integrating Istio service mesh
- 18:20Data correlation and vendor selection
- 23:16Scaling challenges and solutions
- 27:29Current state and OpenTelemetry trends
- 31:42Lessons learned and management insights
- 34:16Q&A session
related talks
49mAdding OpenTelemetry to Production Apps Lessons Learned
26mScaling OpenTelemetry for Modern Workloads: From Mobile To Ma...
05mOpenTelemetry in production for a year plus: the good the bad the ugly
35mOpenTelemetry for Busy People
27mIngesting 6.5 Tb of Telemetry Data Daily Through Open Telemetry Protocol and Collectors
37mHow To Adopt OpenTelemetry in an Enterprise Where Incumbent Vendor Tools Reign Supre...
36mPanel: OpenTelemetry...
34mPanel: OpenTelemetry: Bridging Platform and Enablement