Summaries are distribution metrics optimized for quantiles (e.g., p50, p90, p99). They also support labels for slicing.

​
Create and observe

local Telemetry = Zyntex:Telemetry(10)

local levelTime = Telemetry:Summary("level_completion_seconds", {
    description = "Time to complete the current level"
})

levelTime:observe(47.9, {level="3"})
levelTime:observe(31.3, {level="3"})

​
When to choose summary

  • You primarily care about p50 / p90 / p99 rather than the full histogram.
  • SLA/SLO style monitoring (e.g., “99% of completions under 60s”).
In dashboards, choose a percentile view for Summary metrics to derive p50/p90/p99. If you need the full shape of the distribution, use a Histogram instead.