Sponsored Session: Unifying Observability at DoorDash using Open Stand...
summary
DoorDash replaces an ETL batch pipeline with an OpenTelemetry → Chronosphere pipeline for mobile metrics, cutting volume and cost and improving alert reliability. Mobile counters fail when cumulative metrics fragment across collectors, so the team ships delta metrics end-to-end, avoids temporality conversion and collector staleness windows, and queries raw deltas with sum_over_time + sum_per_second to reconstruct rates reliably for real‑time alerting.
AI-written summary · about
chapters (15)
- 0:00Mobile observability at DoorDash
- 1:05Metrics as observability core
- 3:12Cumulative vs delta temporality
- 5:41Chronosphere control plane
- 6:44Mobile metrics challenges
- 9:15OpenTelemetry for mobile
- 11:00Delta metrics from mobile
- 12:03Temporality conversion issues
- 14:29Cumulative metrics problems
- 16:20End-to-end delta metrics
- 18:31Delta metrics impact on alerting
- 19:15Delta metrics use cases
- 20:49OpenTelemetry delta counters
- 21:53Prometheus delta group
- 23:01Acknowledgments
related talks
26mScaling OpenTelemetry for Modern Workloads: From Mobile To Ma...
21mBuilding Composable OTel Pipelines: CI/CD, Testing, Team-First, and Sca...
35mDoorDash’s Journey From StatsD To Prometheus With 10 Million Metrics/Second
24mOpenTelemetry at Delivery Hero: The Good, the Bad and the Vendor-Agnostic
30mPanel: Telemetry That Matters
45mFrom Noise To Signal: Building Smart Observability Pipelines With OpenTelemetry
34mPanel: OpenTelemetry: Bridging Platform and Enablement
35mMobile Observability at scale