Concepts
-
Sampling: Recording Less Without Knowing Less
How sampling works in monitoring: head vs tail sampling, error-biased strategies, the real odds of missing a bug, and cost math at different rates.
-
Rage Clicks: What Counts, What They Mean, When to Ignore Them
What counts as a rage click, the detection thresholds tools use, the common false positives, and how to tell real frustration from click habits.
-
Breadcrumbs: the Poor Man's Session Replay
What breadcrumbs are in error tracking, what a trail can and cannot tell you, when they're honestly enough, and when you need full session replay.
-
p50, p95, p99: Percentiles for People Who Skipped Stats
What p95 latency means in plain language, why averages hide slow requests, which percentile to alert on, and how to compute them from RUM data.
-
DOM Snapshots: How a Page Becomes Data
What a DOM snapshot is, how a serializer walks the tree, what gets captured versus referenced versus rebuilt, and why replayed pages can't run code.