intermediate · Module 2.5
Grafana Dashboards for OTT
Learning Objectives
- Build RED/USE-style panels for OTT playback services.
- Use template variables to pivot dashboards by CDN PoP and content.
- Add exemplars linking a latency panel to representative traces.
Key Concepts
RED and USE
RED (Rate, Errors, Duration) frames request-driven services like playback; USE (Utilization, Saturation, Errors) frames resources like the Redis buffer. A good OTT dashboard combines both.
Template variables
A $cdn or $content variable rewrites panel queries so one dashboard serves every PoP, avoiding a dashboard-per-PoP sprawl.
Exemplars
Exemplars attach a trace ID to a metric data point, letting you click a latency spike and jump straight into a representative slow trace.
Insight
Design dashboards for the incident you will have at 2am, not for a demo — lead with SLO burn and error ratio, not vanity totals.
Info
Exemplars are the metrics-to-traces bridge and reuse the same trace ID that ties logs together.
Knowledge Check
1. What do the R, E, and D in the RED method stand for?
2. What problem do Grafana template variables solve for an OTT platform?
3. What is the purpose of an exemplar on a latency panel?