Upgrade prometheus, reduce retention time, and drop top metrics.

services/prometheus.hcl

@@ -23,9 +23,14 @@ job "prometheus" {
       driver = "docker"
 
       config {
-        image = "prom/prometheus:v2.44.0"
+        image = "prom/prometheus:v2.52.0"
         network_mode = "host"
-        args = ["--storage.tsdb.path", "/opt/prometheus", "--web.listen-address", "0.0.0.0:9090", "--storage.tsdb.retention.time", "900d"]
+        args = [
+          "--storage.tsdb.path", "/opt/prometheus",
+          "--web.listen-address", "0.0.0.0:9090",
+          "--storage.tsdb.retention.time", "730d",
+          "--web.enable-admin-api"
+        ]
         force_pull = true
         ports = ["http"]
         volumes = [
@@ -98,6 +103,10 @@ scrape_configs:
         regex: '.*job-(.*?)(,.*)'
         replacement: '${1}'
         target_label: 'job_name'
+    metric_relabel_configs:
+      - source_labels: [__name__]
+        regex: '(traefik_service_request_duration_seconds_bucket|traefik_entrypoint_request_duration_seconds_bucket)'
+        action: drop
 
   - job_name: 'consul-server'
     scrape_interval: 10s