Traces

Live observability.

Post-prod telemetry from ask-zeroindex, ingested as it happens. Companion to evals.zeroindex.ai — evals catch regressions before deploy; traces show what actually happens after.

Last 30 days
00 / Traffic

Requests per day

Total events ingested, by day.

01 / Outcomes

Outcome distribution

ok / error / aborted, stacked per day.

02 / Latency

Total response time

p50 / p95 / p99 from request to final byte.

03 / Spend

Cost per day

Estimated USD across input, output, and cached tokens.

04 / Latency

First-token latency

p50 / p95 / p99 from request to first streamed token.

05 / Citations

Citation count distribution

How many sources each answer cites.

06 / Retrieval

Top retrieved chunks

Which content shows up most across results.

Aggregates only — per-event detail is admin-gated.