Sending totals and a daily series — delivery, bounce, and complaint rates over a time range, and how to read them.
One endpoint gives you the numbers that matter for a sending project: totals over a window, a per-day series to chart, and the raw counts you need to compute delivery, bounce, and complaint rates.The metrics endpoint aggregates the same events that drive webhooks — delivery, bounce, complaint, open, click — into totals and a daily breakdown for a project. It returns counts; you compute rates from them, so you stay in control of exactly how each rate is defined.
GET /v1/metricsOptional from and to ISO-8601 timestamps bound the window; omit them for a sensible recent default. The response echoes the resolved range so you always know which window the numbers describe.
# Last 7 days (default range)curl https://api.drin.run/v1/metrics \ -H "Authorization: Bearer $DRIN_API_KEY"# An explicit windowcurl "https://api.drin.run/v1/metrics?from=2026-05-01T00:00:00Z&to=2026-05-31T23:59:59Z" \ -H "Authorization: Bearer $DRIN_API_KEY"
Three fields: the resolved range, a totals map keyed by event type, and a daily array of { date, counts } — one bucket per day, each with its own counts map.
Counts are keyed by event type. Keys in totals and each day’s counts are event types — sent, delivery, bounce, complaint, open, click, and the rest. A type with no events in the window is simply absent, so read defensively (treat a missing key as 0).
Rates are ratios over sent. Compute them yourself from totals so the denominator is unambiguous:
const t = metrics.totals;const sent = t.sent ?? 0;const deliveryRate = sent ? (t.delivery ?? 0) / sent : 0; // want: highconst bounceRate = sent ? (t.bounce ?? 0) / sent : 0; // want: < 4%const complaintRate = sent ? (t.complaint ?? 0) / sent : 0; // want: < 0.1%const openRate = sent ? (t.open ?? 0) / sent : 0;const clickRate = sent ? (t.click ?? 0) / sent : 0;console.log((bounceRate * 100).toFixed(2) + "% bounce");
Delivery rate — delivery / sent. The share that reached the mailbox provider. Higher is better.
Bounce rate — bounce / sent. The single most important number to watch. Keep it well under 4%.
Complaint rate — complaint / sent. Spam-button presses. Keep it under 0.1%.
Open / click rate — engagement signals. Useful trends, but treat the absolutes with care (privacy proxies inflate opens; not every client loads tracking).
The daily series is built for time-series charts. Map each bucket the same way you mapped the totals to get a rate-over-time line:
// Plot a bounce-rate-over-time line from the daily series.const points = metrics.daily.map((d) => { const sent = d.counts.sent ?? 0; return { date: d.date, bounceRate: sent ? (d.counts.bounce ?? 0) / sent : 0 };});
Watch the trend, not a single day. A one-day spike on low volume is noise; a rising multi-day trend is a signal. Plot the bounce and complaint lines against your thresholds and act on sustained movement.
The lines that matter. Drin watches the same limits the mailbox providers enforce. Sustained bounce above 4% or complaints above 0.1% put your sending — and eventually the whole platform’s — reputation at risk, and can trigger an automatic pause. If a rate is climbing, slow down, verify your list, and check the Deliverability guide before it trips.
High bounce usually means a stale or unverified recipient list — clean it and lean on suppressions. High complaints usually mean unwanted mail — tighten consent and make unsubscribing easy. The metrics endpoint is your early-warning system; check it before a provider does it for you.
