Synthetic probe (internal)
Meridian runs synthetic probes every 60 seconds against every deployed region. Each probe simulates a full end-to-end transaction — DNS resolve, TLS handshake, HTTP request, and response validation — and reports latency, status code, and body checksum back to the control plane.
Probe lifecycle
- Dispatch — the scheduler fans out probe jobs to edge workers in every active region simultaneously.
- Execute — each worker performs the full transaction and records per-phase timing (DNS, TCP, TLS, TTFB, download).
- Validate — the response body is hashed and compared against a known golden checksum. Mismatches trigger an immediate alert.
- Report — results are written to the metrics store and surfaced on the dashboard within seconds.
Alerting
When a probe fails — timeout, non-2xx status, checksum mismatch, or region unreachable — Meridian fires a Discord webhook into the ops channel. The alert payload includes the failing region, the specific phase that broke, and a link to the probe trace. Consecutive failures escalate the mention level.
Regions
Synthetic probes are internal infrastructure. They are not exposed to customer accounts and do not count against your monitor quota. Every Meridian plan includes full synthetic coverage at no extra cost.