CoreDNS-GSLB: Observability

Prometheus Metrics

If you enable the prometheus block in your Corefile, the plugin exposes the following metrics on /metrics (default port 9153):

Enable in Corefile:

. {
    prometheus
    ...
}

Available metrics:

Metric Name	Labels	Description
`gslb_healthcheck_total`	`name`, `type`, `address`, `result`	Total number of healthchecks performed.
`gslb_healthcheck_duration_seconds`	`type`, `address`	Duration of healthchecks in seconds.
`gslb_healthcheck_failures_total`	`type`, `address`, `reason`	Total number of healthcheck failures. `reason` can be: `timeout`, `connection`, `protocol`, `other`.
`gslb_record_resolution_total`	`name`, `result`	Total number of GSLB record resolutions.
`gslb_record_resolution_duration_seconds`	`name`, `result`	Duration of GSLB record resolution in seconds.
`gslb_record_health_status`	`name`	Health status per record (1 = healthy, 0 = unhealthy).
`gslb_backend_health_status`	`name`, `address`	Health status per backend (2 = disabled, 1 = healthy, 0 = unhealthy).
`gslb_backend_healthcheck_status`	`name`, `address`, `type`	Healthcheck status per backend and type (2 = disabled, 1 = success, 0 = fail).
`gslb_config_reload_total`	`result`	Total number of config reloads.
`gslb_backend_active`	`name`	Number of active (healthy) backends per record.
`gslb_backend_selected_total`	`name`, `address`	Total number of times a backend was selected for a record.
`gslb_healthchecks_total`	(none)	Number of healthchecks configured (total for all records/backends).
`gslb_backends_total`	(none)	Total number of backends configured (all records).
`gslb_records_total`	(none)	Total number of GSLB records configured.
`gslb_zones_total`	(none)	Total number of DNS zones configured.
`gslb_version_info`	`version`	GSLB build version info (always set to 1).
`gslb_redis_connected`	(none)	Connection status to Redis (1 = connected, 0 = disconnected).
`gslb_redis_errors_total`	`operation`	Total number of Redis errors labeled by operation.
`gslb_redis_operations_total`	`operation`, `result`	Total number of Redis operations labeled by operation and result.

You can then scrape metrics at http://localhost:9153/metrics

Grafana Dashboard

The dashboard is available in dashboards/gslb-observability.json

CoreDNS-GSLB

Using the Simplified Health Status Metrics

gslb_record_health_status{name="..."}: 1 if at least one backend is healthy, 0 if all are unhealthy or disabled.
gslb_backend_health_status{name="...", address="..."}: 2 if backend is disabled, 1 if healthy, 0 if unhealthy.

Example Prometheus Queries

Count unhealthy backends:
```
sum(gslb_backend_health_status == 0)
```

Show 0 when all backends are healthy/disabled:

sum(gslb_backend_health_status == 0) or on() vector(0)

Count healthy records:
```
sum(gslb_record_health_status)
```
Count unhealthy records:
```
sum(1 - gslb_record_health_status)
```

These queries ensure that Grafana panels always display 0 when everything is healthy, instead of showing nothing.