← WORK
WORK / 2026 PLATFORM BUILDING

VAST — Monitoring & Alerting

Central monitoring and alerting (Go) — the built-in health and alerting layer for WAAP, branded VAST. Probes endpoints, watches WAF/WAAP signals, and fires on fixed thresholds or statistical anomaly detection.

GoTimescaleDBRedisElastic

Why

WAAP and the edge fleet need one place that knows whether services are healthy and tells the right people when they aren’t — without bolting on a third-party monitoring stack. VAST is that place: the built-in health and alerting layer that WAAP owns, rather than an external stack glued on after the fact.

What it does

VAST watches several signal sources and decides, for each, what counts as a failure worth telling someone about.

  • Probes uptime and latency for HTTP/HTTPS/TCP/DNS/TLS endpoints.
  • Watches security and traffic signals — WAF logs from Elastic, and WAAP / delivery analytics.
  • Evaluates alarm conditions two ways. A fixed threshold fires when a value crosses a preset line. Anomaly detection ($sigma family) instead fires when a value departs from what is normal for that endpoint — measured against a rolling 7-day baseline with per-hour seasonal stats, so the line that counts as “failure” moves with time of day rather than staying fixed.
  • Notifies subscribers over Email, Telegram, Slack, Teams and webhooks, with an anomaly chart attached.

Architecture

G watchpoint watchpoint probe probe engine watchpoint->probe leader-elected via Redis lease handler result handler probe->handler alerter alerter handler->alerter queue for async alarm evaluation tsdb TimescaleDB handler->tsdb persist dispatcher dispatcher alerter->dispatcher subscribers subscribers dispatcher->subscribers baseline baseline job (hourly, 7-day rolling) tsdb->baseline recompute hourly baseline->alerter

One replica is leader-elected to run the probe engine and alerting; the rest serve the API. A firing alarm — whether a crossed fixed threshold or a flagged anomaly — becomes an incident that follows a triggered → acknowledged → resolved lifecycle, with cooldown-based re-notification and auto-resolve on recovery.

Status

In active development — the built-in monitoring for WAAP.