API RequestsiDistinct task_id count — one per OpenAPI scraping request. Active UsersiUnique user_id values that called the OpenAPI. Active API KeysiUnique api_key values used in the range. Success RateiSUCCESS / total stage events. Avg LLM ExtractiAverage duration_ms of successful LLM_EXTRACT stage events. Failed EventsiCount of stage events with event=FAIL. API Request VolumeiCount of OpenAPI scraping requests over time (one per HTML_FETCH event).
API Type DistributioniBreakdown of requests by api_type (DISTILL vs EXTRACT).
Active Users / API Keys Over TimeiDistinct user_id and api_key counts per time bucket (one HTML_FETCH event per request).
Daily New UsersiUsers whose first-ever OpenAPI request (within a 90d lookback) landed on each day. X-axis follows the selected time range.
Requests by API TypeiPer-api_type request count over time.
Top API KeysiOne row per API key — same customer can hold multiple keys. Email resolved from user_id via internal tooling.
Top UsersiTop user_id values by OpenAPI request volume, resolved to customer email.
Top Scraped DomainsiDomains most frequently scraped via the OpenAPI.
Errors by Class & StageiFailure events grouped by error_class and pipeline stage.
Stage Performance (Successful Events)iLatency percentiles per pipeline stage for SUCCESS events.
Monitor