<observations>
Date: May 1, 2026
* 🟡 (05:45) User suggested converting "Resources Provisioned" and "Deep Link Logins" widgets to big_number type
* 🟡 (05:45) User noted numbers across dashboard are very low except for ~8k spike on "Stripe Projects Activity by Endpoint" widget — asked if numbers make sense
* 🟡 (05:46) Agent queried spans dataset for 7-day counts by endpoint: account_requests: 7,725; resources: 1,423; services: 2,381; deep-link-login: 770
* 🟡 (05:46) Agent found account_requests HTTP breakdown: 6,695 with HTTP None (HMAC failures rejected before view), 1,030 with HTTP 200
* 🟡 (05:46) Agent found resources endpoint HTTP breakdown: 1,684 with HTTP None, 696 with HTTP 200
* 🟡 (05:47) Agent found only 2 log event types reaching ourlogs in 30 days: `stripe_projects.deep_link.created` (128 total) and `stripe_projects.resource.service_updated` (64 total) — out of 60+ distinct event names the code emits
* 🟡 (05:47) Root cause analysis: 8k spike is real but inflated by HMAC permission check failures (6,695 requests with no HTTP status); services/ endpoint ~100/day from Stripe's 10-minute polling is expected behavior
* 🟡 (05:47) Root cause analysis: near-zero log counts are a logging pipeline issue — `logger = logging.getLogger("getsentry.stripe_projects")` likely not hooked into Sentry SDK's ourlogs ingestion pipeline; all logs-based dashboard widgets are underreporting
* 🟡 (05:48) Agent converted "Resources Provisioned" and "Deep Link Logins" widgets to big_number type on dashboard 4510068
* 🟡 (05:48) Agent flagged the logging gap as the bigger issue — most `logger.info(...)` calls in `getsentry/web/stripe_projects/` not reaching ourlogs; offered to investigate logging configuration
</observations>