Metrics

Metrics are exposed in Prometheus text format at /metrics path.

NameTypeDescriptionLabels
retry_request_totalCounterThe number of retry requestsuser, cluster, cluster_user, replica, cluster_node
bad_requests_totalCounterThe number of unsupported requests
cache_hits_totalCounterThe amount of cache hitscache, user, cluster, cluster_user
cache_itemsGaugeThe number of items in each cachecache
cache_miss_totalCounterThe amount of cache missescache, user, cluster, cluster_user
cache_sizeGaugeSize of each cachecache
cached_response_duration_secondsSummaryDuration for cached responses. Includes the duration for sending response to clientcache, user, cluster, cluster_user
canceled_request_totalCounterThe number of requests canceled by remote clientuser, cluster, cluster_user, replica, cluster_node
cluster_user_queue_overflow_totalCounterThe number of overflows for per-cluster_user request queuesuser, cluster, cluster_user
concurrent_limit_excess_totalCounterThe number of rejected requests due to max_concurrent_queries limituser, cluster, cluster_user, replica, cluster_node
concurrent_queriesGaugeThe number of concurrent queries at the momentuser, cluster, cluster_user, replica, cluster_node
config_last_reload_successfulGaugeWhether the last configuration reload attempt was successful
config_last_reload_success_timestamp_secondsGaugeTimestamp of the last successful configuration reload
host_healthGaugeHealth state of hosts by clusterscluster, replica, cluster_node
host_penalties_totalCounterThe number of given penalties by hostcluster, replica, cluster_node
killed_request_totalCounterThe number of requests killed by proxyuser, cluster, cluster_user, replica, cluster_node
proxied_response_duration_secondsSummaryDuration for responses proxied from ClickHouseuser, cluster, cluster_user, replica, cluster_node
request_body_bytes_totalCounterThe amount of bytes read from request bodiesuser, cluster, cluster_user, replica, cluster_node
request_duration_secondsSummaryRequest duration. Includes possible queue wait timeuser, cluster, cluster_user, replica, cluster_node
request_queue_sizeGaugeRequest queue size at the momentuser, cluster, cluster_user
request_success_totalCounterThe number of successfully proxied requestsuser, cluster, cluster_user, replica, cluster_node
request_sum_totalCounterThe number of processed requestsuser, cluster, cluster_user, replica, cluster_node
response_body_bytes_totalCounterThe amount of bytes written to response bodiesuser, cluster, cluster_user, replica, cluster_node
status_codes_totalCounterDistribution by response status codesuser, cluster, cluster_user, replica, cluster_node, code
timeout_request_totalCounterThe number of timed out requestsuser, cluster, cluster_user, replica, cluster_node
user_queue_overflow_totalCounterThe number of overflows for per-user request queuesuser, cluster, cluster_user

An example of Grafana's dashboard for chproxy metrics is available here.

dashboard example

Edit this page on GitHub Updated at Tue, Nov 29, 2022