Observability
How well you can understand what a system is doing from the data it emits, usually logs, metrics, and traces. Good observability lets you answer new questions about a failure without shipping new code.
Observability
In Quorum#
Observability Research shows you more of the Ledger. Structured logs split errors by cause, Metrics adds history graphs to every Node, Percentiles shows p50, p95, and p99, and Alerting raises a pager banner when errors, p99, or saturation cross a threshold. The Analytics Node samples the traffic of Nodes that you link into it, from 1% to 100%, and turns it into Insight. Each sample teaches more as Apdex falls, up to twice as much, and a higher sample rate costs more CPU and Burn. Metrics also opens the Dashboard (D), which charts your Design over time.
Read next#
Distributed tracing, Percentile (p50, p95, p99), SLO, SLA, SLI, MTTR