Failover, weighted, geo, and latency routing — all four, all at the same plan. Real-time propagation, a clean API, and an admin UI built for engineers.
Most providers serve auth DNS. The differentiators are how the answers get picked, how fast changes propagate, and how nice the day-to-day is.
Health-check your endpoints. If the primary stops responding, the engine serves the backup automatically.
70/30 between two A records, or 50/30/20 across three. Bias new versions, run canaries, do active-active.
EU users hit your EU stack, NA users hit your NA stack. ECS (EDNS Client Subnet) honored when present.
Each PoP returns the target with the lowest probe latency from its own perspective. Closer is better.
No "wait 24 hours for DNS to propagate." Changes hit every engine within milliseconds via NATS — TTL still gates downstream resolver caches, of course.
Click "Test query" on any record — dnswiz fires a real DNS query against your engine and shows the answer, RTT, and which GSLB target won and why.
Endpoint goes down at 3am? dnswiz posts a signed JSON payload to your Slack / PagerDuty / custom URL. You know before your customers do.
All plans include all four GSLB types. The only thing plans gate is volume.
Go + Postgres + NATS + JetStream. Open API, OpenAPI spec, Terraform provider. No surprise SaaS lock-in.