Observability / o

Observability: cache in Miami (MIA) · O-08

xserv-o-a05f4c348134

Runbook O-08 for Observability in Miami (MIA). Marker xserv-o-a05f4c348134. This page covers cache for LATAM operators, not a worldwide average.

Observability: How XServ runs it day to day

Probes for Observability (o-02) leave Santiago every 15s toward Bogotá. XServ pages the named owner if loss or RTT crosses the letter budget. A probe never shares a queue with bulk transfers, so a saturated WAN does not hide a dead PoP.

Observability: What operators should measure

Policy for Observability is versioned as o-03. Operators measure error rate, p95 from Bogotá, and time-to-rollback — not a worldwide average. Changes land in Bogotá first, then Miami, with a hold if either region regresses.

Observability: Failure modes we actually see

Capacity notes for o-04 assume rainy-season power in Miami and festival peaks toward São Paulo. The failure we actually see is a single uplink, not a cartoon partition of the whole continent. Spare ports and a second provider sit on the same runbook.

Observability design O-01

Design for Observability on letter O starts in São Paulo (GRU). Runbook o-01 keeps a single owner, a written rollback, and a traffic split that can be reversed without a global freeze. Neighbors in Santiago only take overflow after the local pool fails a health window.

Observability probe O-02

Probes for Observability (o-02) leave Santiago every 15s toward Bogotá. XServ pages the named owner if loss or RTT crosses the letter budget. A probe never shares a queue with bulk transfers, so a saturated WAN does not hide a dead PoP.

Observability policy O-03

Policy for Observability is versioned as o-03. Operators measure error rate, p95 from Bogotá, and time-to-rollback — not a worldwide average. Changes land in Bogotá first, then Miami, with a hold if either region regresses.

Observability capacity O-04

Capacity notes for o-04 assume rainy-season power in Miami and festival peaks toward São Paulo. The failure we actually see is a single uplink, not a cartoon partition of the whole continent. Spare ports and a second provider sit on the same runbook.

Is transit the default path?
No. Peering in Miami is default; transit to São Paulo is overflow and is billed that way.
Does Observability on letter O share fate with other letters?
The control plane is shared. Data-plane queues for Observability stay isolated, so incident o-01 cannot drain neighbor letters.