fix(monitoring): use active_workers metric instead of kube_job query

CI training jobs run as GitLab runner pods, not K8s Jobs, so the
kube_job_status_active query always returned 0. Switch to
foxhunt_training_active_workers which is emitted by all training
binaries regardless of deployment method.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

services/monitoring_service/src/prometheus_client.rs

@@ -89,10 +89,10 @@ impl PrometheusClient {
         Ok(parse_samples(results))
     }
 
-    /// Fetch active K8s training job count
+    /// Fetch active training worker count (covers both K8s Jobs and CI runner pods)
     pub async fn fetch_active_jobs(&self) -> Result<u32> {
         let results = self
-            .query(r#"kube_job_status_active{namespace="foxhunt",job_name=~"training-.*"}"#)
+            .query("foxhunt_training_active_workers")
             .await?;
         let count: f64 = results
             .iter()