Mesh Observability

Real-time view of the ParlayAPI ingest layer. Per-node telemetry, per-book ingest status, per-ISP reachability. Counts are observed numbers; no estimates.

Loading...

Mesh aggregate

Per-node telemetry

Every ingest node currently contributing data to this worker. Node ids are class-prefixed (residential / edge / cloud) without naming the underlying host vendor. Status thresholds: active <3min, active_recent <30min, stalled_minor <2h, stalled_major <6h, offline ≥6h.

Node Status Last post Posts 1h Posts 24h Bytes 24h Sources
Loading per-node telemetry...

Per-book ingest stack

For every upstream sportsbook we track: the bot-protection vendor we observed, our routing strategy, observed end-to-end refresh, and any genuine limitations.

Book Region Bot protection Ingest path Refresh Limitations
Loading per-book stack...

ISP reachability snapshot

Of the 9 major US residential and business ISPs we track, here's which have reached this worker recently. Counts are exact observed requests since worker boot. Visit /diag to debug your own connection if your ISP is on the filter list.

ISP Status Last seen Requests 24h Filter history
Loading ISP reachability...
How this works. Three public endpoints power this dashboard: /v1/meta/mesh-status for per-node telemetry, /v1/meta/source-stack for the per-book stack details, and /v1/meta/isp-reachability for the ISP-level reachability roll-up. All three are free, public, no-auth. Polled every 30 seconds from this page. The data is exact: observed counts since the current worker booted, never extrapolated. Counts reset on rolling deploy.

Worker boot times: multiple worker processes serve the load balancer, so counts on this page reflect one worker's view per refresh. Reload to round-robin to another worker for the cluster shape.