prometheus: use apiers_conns for StatQueue fetch

agents/prometheus.go

@@ -21,7 +21,6 @@ package agents
 import (
 	"fmt"
 	"net/http"
-	"strings"
 
 	"github.com/cgrates/birpc/context"
 	"github.com/cgrates/cgrates/config"
@@ -217,20 +216,12 @@ func (pa *PrometheusAgent) ServeHTTP(w http.ResponseWriter, r *http.Request) {
 // updateStatsMetrics fetches and updates all StatQueue metrics by calling each
 // configured StatS connection.
 func (pa *PrometheusAgent) updateStatsMetrics() {
-	for _, connID := range pa.cfg.PrometheusAgentCfg().StatSConns {
+	for connIdx, connID := range pa.cfg.PrometheusAgentCfg().StatSConns {
 		sqIDs := pa.cfg.PrometheusAgentCfg().StatQueueIDs
 
 		// When no StatQueueIDs set, fetch all available ones.
 		if len(sqIDs) == 0 {
-
-			// Internal StatS connections cannot handle APIerS calls.
-			// Redirect *internal:*stats to *internal:*apier to get StatQueue IDs.
-			apiersConnID := connID
-			if strings.HasPrefix(connID, utils.MetaInternal) {
-				apiersConnID = utils.ConcatenatedKey(utils.MetaInternal,
-					utils.MetaApier)
-			}
-
+			apiersConnID := pa.cfg.PrometheusAgentCfg().ApierSConns[connIdx]
 			if err := pa.cm.Call(context.Background(), []string{apiersConnID},
 				utils.APIerSv1GetStatQueueProfileIDs,
 				&utils.PaginatorWithTenant{}, &sqIDs); err != nil {

config/config_defaults.go

@@ -841,6 +841,7 @@ const CGRATES_CFG_JSON = `
 "prometheus_agent": {
 	"enabled": false,			// enables the prometheus agent: <true|false>
 	"path": "/prometheus",			// endpoint for prometheus metrics
+	"apiers_conns": [],			// connections to ApierS, empty to disable: <""|*internal|$rpc_conns_id>
 	"caches_conns": [], 			// connections to CacheS, empty to disable: <""|*internal|$rpc_conns_id>
 	"cache_ids": [],			// cache partition IDs to collect statistics for, empty for all partitions
 	"cores_conns": [],			// connections to CoreS, empty to disable: <""|*internal|$rpc_conns_id>

config/configsanity.go

@@ -1341,10 +1341,18 @@ func (cfg *CGRConfig) checkConfigSanity() error {
 	if cfg.prometheusAgentCfg.Enabled {
 		if len(cfg.prometheusAgentCfg.StatSConns) > 0 &&
 			len(cfg.prometheusAgentCfg.StatQueueIDs) == 0 &&
-			!cfg.apier.Enabled {
+			len(cfg.prometheusAgentCfg.StatSConns) != len(cfg.prometheusAgentCfg.ApierSConns) {
 			return fmt.Errorf(
-				"<%s> when StatQueueIDs is empty, %s must be enabled to retrieve all available StatQueue IDs",
-				utils.PrometheusAgent, utils.ApierS)
+				"<%s> when StatQueueIDs is empty, apiers_conns must match stats_conns length to fetch StatQueue IDs",
+				utils.PrometheusAgent)
+		}
+		for _, connID := range cfg.prometheusAgentCfg.ApierSConns {
+			if strings.HasPrefix(connID, utils.MetaInternal) && !cfg.apier.Enabled {
+				return fmt.Errorf("<%s> not enabled but requested by <%s> component", utils.ApierS, utils.PrometheusAgent)
+			}
+			if _, has := cfg.rpcConns[connID]; !has && !strings.HasPrefix(connID, utils.MetaInternal) {
+				return fmt.Errorf("<%s> connection with id: <%s> not defined", utils.PrometheusAgent, connID)
+			}
 		}
 		for _, connID := range cfg.prometheusAgentCfg.StatSConns {
 			if strings.HasPrefix(connID, utils.MetaInternal) && !cfg.statsCfg.Enabled {

config/prometheus.go

@@ -33,6 +33,7 @@ type PrometheusAgentJsonCfg struct {
 	CacheSConns           *[]string `json:"caches_conns"`
 	CacheIDs              *[]string `json:"cache_ids"`
 	CoreSConns            *[]string `json:"cores_conns"`
+	ApierSConns           *[]string `json:"apiers_conns"`
 	StatSConns            *[]string `json:"stats_conns"`
 	StatQueueIDs          *[]string `json:"stat_queue_ids"`
 }
@@ -46,6 +47,7 @@ type PrometheusAgentCfg struct {
 	CacheSConns           []string
 	CacheIDs              []string
 	CoreSConns            []string
+	ApierSConns           []string
 	StatSConns            []string
 	StatQueueIDs          []string
 }
@@ -75,6 +77,9 @@ func (c *PrometheusAgentCfg) loadFromJSONCfg(jc *PrometheusAgentJsonCfg) error {
 	if jc.CoreSConns != nil {
 		c.CoreSConns = tagInternalConns(*jc.CoreSConns, utils.MetaCore)
 	}
+	if jc.ApierSConns != nil {
+		c.ApierSConns = tagInternalConns(*jc.ApierSConns, utils.MetaApier)
+	}
 	if jc.StatSConns != nil {
 		c.StatSConns = tagInternalConns(*jc.StatSConns, utils.MetaStats)
 	}
@@ -94,6 +99,7 @@ func (c PrometheusAgentCfg) AsMapInterface() any {
 		utils.CacheSConnsCfg:           stripInternalConns(c.CacheSConns),
 		utils.CacheIDsCfg:              stripInternalConns(c.CacheIDs),
 		utils.CoreSConnsCfg:            stripInternalConns(c.CoreSConns),
+		utils.ApierSConnsCfg:           stripInternalConns(c.ApierSConns),
 		utils.StatSConnsCfg:            stripInternalConns(c.StatSConns),
 		utils.StatQueueIDsCfg:          c.StatQueueIDs,
 	}
@@ -109,6 +115,7 @@ func (c PrometheusAgentCfg) Clone() *PrometheusAgentCfg {
 		CacheSConns:           slices.Clone(c.CacheSConns),
 		CacheIDs:              slices.Clone(c.CacheIDs),
 		CoreSConns:            slices.Clone(c.CoreSConns),
+		ApierSConns:           slices.Clone(c.ApierSConns),
 		StatSConns:            slices.Clone(c.StatSConns),
 		StatQueueIDs:          slices.Clone(c.StatQueueIDs),
 	}

docs/prometheus.rst

@@ -21,6 +21,7 @@ Example configuration in the JSON file:
     "prometheus_agent": {
         "enabled": true,
         "path": "/prometheus",
+        "apiers_conns": ["*internal", "external"],
         "caches_conns": ["*internal"],
         "cache_ids": [
             "*attribute_filter_indexes",
@@ -43,6 +44,9 @@ enabled
 path
     HTTP endpoint path where Prometheus metrics will be exposed, e.g., "/prometheus" or "/metrics"
 
+apiers_conns
+    List of connection IDs to ApierS components. Required when stat_queue_ids is empty to fetch all available StatQueue profile IDs. Must match the length of stats_conns when auto-fetching is used. Possible values: <""|*internal|$rpc_conns_id>
+
 caches_conns
     List of connection IDs to CacheS components for collecting cache statistics. Empty list disables cache metrics collection. Possible values: <""|*internal|$rpc_conns_id>
 
@@ -56,7 +60,7 @@ stats_conns
     List of connection IDs to StatS components for collecting StatQueue metrics. Empty list disables StatQueue metrics collection. Possible values: <""|*internal|$rpc_conns_id>
 
 stat_queue_ids
-    List of StatQueue IDs to collect metrics from. Can include tenant in format <[tenant]:ID>. If tenant is not specified, default tenant from general configuration is used.
+    List of StatQueue IDs to collect metrics from. Can include tenant in format <[tenant]:ID>. If tenant is not specified, default tenant from general configuration is used. Leave empty to automatically collect metrics from all available StatQueues (requires apiers_conns).
 
 Available Metrics
 -----------------
@@ -135,6 +139,11 @@ The PrometheusAgent operates differently than other CGRateS components that use
 - When multiple connections are configured in caches_conns, the agent collects cache statistics from **all** connections for the specified cache_ids
 - The agent processes metrics requests only when Prometheus sends a scrape request to the configured HTTP endpoint
 
+StatQueue metrics are collected based on the ``stat_queue_ids`` configuration. When specific StatQueue IDs are provided, only those StatQueues are monitored. When ``stat_queue_ids`` is left empty, all available StatQueues are monitored by fetching StatQueue profile IDs from the configured ``apiers_conns``.
+
+.. note::
+    When fetching all StatQueues (empty stat_queue_ids), each ApierS connection in ``apiers_conns`` corresponds to its StatS counterpart at the same index position in ``stats_conns``.
+
 You can view all exported metrics and see what Prometheus would scrape by making a simple curl request to the HTTP endpoint:
 
 .. code-block:: bash

general_tests/prometheus_it_test.go

@@ -67,6 +67,7 @@ func TestPrometheusAgentIT(t *testing.T) {
 		"*stat_filter_indexes",
 		"*rpc_connections"
 	],
+	// "apiers_conns": ["*internal", "external"],
 	"stats_conns": ["*internal", "external"],
 	"stat_queue_ids": ["cgrates.org:SQ_1","SQ_2"]
 }