Monitoring

Labyrinth observes itself three ways: Prometheus metrics, a Zabbix agent, and structured logging with a WebSocket query stream.

Prometheus

Scrape the metrics endpoint:

http://labyrinth:9153/metrics

Counters, histograms and gauges include:

labyrinth_queries_total{type="A"} 28041
labyrinth_cache_hits_total 18294
labyrinth_cache_misses_total 9747
labyrinth_query_duration_seconds_bucket{le="0.001"} 15000
labyrinth_uptime_seconds 86400

Zabbix

Two integration modes:

HTTP agent (recommended) — configure Zabbix items to poll:

http://labyrinth:9153/api/zabbix/item?key=labyrinth.cache.hit_ratio

Native agent — enable zabbix.addr in the config for the native Zabbix agent protocol on TCP port 10050.

Available keys:

labyrinth.queries.total
labyrinth.cache.hits
labyrinth.cache.misses
labyrinth.cache.hit_ratio
labyrinth.cache.entries
labyrinth.upstream.queries
labyrinth.upstream.errors
labyrinth.uptime
labyrinth.goroutines

Health checks

Logs

Structured JSON or text logging via logging.format — see Configuration. The dashboard’s Operations page surfaces reliability incidents from the same counters.

Source: the project README.