A lightweight monitoring agent you run on your own compute (a VM, container, or Kubernetes) in the regions you care about. It measures DNS resolution latency and target reachability for your Stackryze zones from that vantage point, then pushes only aggregated metrics to the Stackryze control plane. The analysis UI lives in the Stackryze dashboard (dns.stackryze.com).
your region (VM / pod)
│
stackryze-edge ──resolve + dial──► your zones' records
│ aggregates locally (latency p50/p95, success %, targets up/down)
▼ Bearer token
api-dns.stackryze.com/api/edge/metrics
│
▼
Stackryze dashboard → Edge monitoring
No raw queries ever leave your infrastructure — only per-window aggregates. Runs entirely on your compute, so it never touches Stackryze's serving path.
| Variable | Required | Default | Description |
|---|---|---|---|
STACKRYZE_API_TOKEN |
yes | — | Token with write scope (Settings → API tokens) |
STACKRYZE_API_URL |
no | https://api-dns.stackryze.com/api |
API base |
STACKRYZE_EDGE_REGION |
no | default |
Vantage-point label shown in the UI (e.g. mumbai) |
STACKRYZE_EDGE_AGENT_ID |
no | <region>-<hostname> |
Stable agent identifier |
STACKRYZE_EDGE_INTERVAL |
no | 60 |
Seconds between cycles (min 15) |
STACKRYZE_EDGE_ZONES |
no | (all your zones) | Comma list to monitor; empty = auto-discover |
STACKRYZE_EDGE_RESOLVERS |
no | 8.8.8.8,1.1.1.1,9.9.9.9 |
Resolver IPs to test against |
STACKRYZE_API_TOKEN=sk_dns_xxx \ STACKRYZE_EDGE_REGION=mumbai \ ./stackryze-edge
docker run -d --name stackryze-edge \ -e STACKRYZE_API_TOKEN=sk_dns_xxx \ -e STACKRYZE_EDGE_REGION=mumbai \ ghcr.io/stackryze/stackryze-edge:latest
apiVersion: apps/v1 kind: Deployment metadata: name: stackryze-edge spec: replicas: 1 selector: matchLabels: { app: stackryze-edge } template: metadata: labels: { app: stackryze-edge } spec: containers: - name: agent image: ghcr.io/stackryze/stackryze-edge:latest env: - name: STACKRYZE_EDGE_REGION value: mumbai - name: STACKRYZE_API_TOKEN valueFrom: secretKeyRef: { name: stackryze-dns, key: token }
Per zone, per cycle:
- Resolution latency (p50/p95) resolving the apex across the configured resolvers.
- Success rate of those lookups.
- Target reachability — dials each resolved IP on
:443(up/down counts).
Requires Go 1.22+. Pure standard library, no external dependencies.
go build -o stackryze-edge .