← All articles

DOC_001 · 29 Oct 2026 · 2 MIN READ

Automated DNS failover with a small Go daemon


How it works

For each configured record, the daemon checks the primary health URL on an interval. After a number of consecutive failures it looks at the secondary: if that one is healthy it moves the DNS record to the secondary IP, otherwise it raises an alert because there is nowhere to go.

How the daemon decides: check, fail over, restore

For each configured recordCheck primaryevery 30sOKKO × 3 consecutiveState NORMALcounter resetCheck secondaryafter 3 consecutive KOsecondary OKsecondary KOSwitch DNS to secondary+ DiscordDiscord alertnowhere to goFAILOVER modekeep checking the primaryOK for 5 stable minutesRestore DNS to primary+ Discordback to NORMALCheck secondaryin parallelKO × 5 consecutiveDiscord alertdegraded fallback

Configuration

Everything lives in one YAML file: the Cloudflare token and zone, the Discord webhook, the thresholds and the list of records to supervise.

cf_api_token: "your_cloudflare_token"
cf_zone_id: "your_zone_id"
discord_webhook_url: "https://discord.com/api/webhooks/xxx/yyy"

check_interval: "30s"
health_check_timeout: "5s"
failover_threshold: 3
secondary_alert_threshold: 5
stability_duration: "5m"
dry_run: false

records:
  - name: "api"
    cf_record_name: "api.example.com"
    primary_ip: "1.2.3.4"
    primary_health_url: "https://api.example.com/health"
    secondary_ip: "5.6.7.8"
    secondary_health_url: "https://api-secondary.example.com/health"
curl -s -X GET "https://api.cloudflare.com/client/v4/user/tokens/verify" \
  -H "Authorization: Bearer YOUR_TOKEN" | jq .success
# must return: true

Deploying

The daemon is compiled on the server and run as a systemd service.

go build -o dns-failover ./
sudo useradd -r -s /bin/false dns-failover
sudo cp dns-failover /opt/dns-failover/
sudo cp config.yaml /opt/dns-failover/config.yaml
sudo cp dns-failover.service /etc/systemd/system/
sudo systemctl enable --now dns-failover
journalctl -u dns-failover -f

Rehearsing without risk