From b13a651cfe979f117466811f334e9fc558b696c5 Mon Sep 17 00:00:00 2001
From: Dalton Hubble <dghubble@gmail.com>
Date: Sat, 9 Feb 2019 16:47:19 -0800
Subject: [PATCH] Drop metrics that are unset, high cardinality, or extraneous

* https://github.com/coreos/prometheus-operator/pull/2387
* https://github.com/coreos/prometheus-operator/pull/1959
---
 CHANGES.md                    |  2 +-
 addons/prometheus/config.yaml | 23 +++++++++++++++++++++++
 2 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/CHANGES.md b/CHANGES.md
index f89ff56f..e2cb82c1 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -10,7 +10,7 @@ Notable changes between versions.
 #### Addons
 
 * Raise nginx-ingress liveness/readiness timeout to 5 seconds
-* Improve Prometheus metrics labels
+* Improve Prometheus metrics labels and drop extraneous metrics ([#397](https://github.com/poseidon/typhoon/pull/397))
   * Add `pod` name label to metrics discovered via service endpoints
   * Rename `kubernetes_namespace` label to `namespace`
 
diff --git a/addons/prometheus/config.yaml b/addons/prometheus/config.yaml
index 2408b3de..3ea5b65d 100644
--- a/addons/prometheus/config.yaml
+++ b/addons/prometheus/config.yaml
@@ -55,6 +55,17 @@ data:
         action: replace
         target_label: job
 
+      metric_relabel_configs:
+      - source_labels: [__name__]
+        action: drop
+        regex: etcd_(debugging|disk|request|server).*
+      - source_labels: [__name__]
+        action: drop
+        regex: apiserver_admission_controller_admission_latencies_seconds_.*
+      - source_labels: [__name__]
+        action: drop
+        regex: apiserver_admission_step_admission_latencies_seconds_.*
+
     # Scrape config for node (i.e. kubelet) /metrics (e.g. 'kubelet_'). Explore
     # metrics from a node by scraping kubelet (127.0.0.1:10250/metrics).
     - job_name: 'kubelet'
@@ -89,6 +100,13 @@ data:
       relabel_configs:
       - action: labelmap
         regex: __meta_kubernetes_node_label_(.+)
+      metric_relabel_configs:
+      - source_labels: [__name__, image]
+        action: drop
+        regex: container_([a-z_]+);
+      - source_labels: [__name__]
+        action: drop
+        regex: container_(network_tcp_usage_total|network_udp_usage_total|tasks_state|cpu_load_average_10s)
 
 
     # Scrap etcd metrics from controllers via listen-metrics-urls
@@ -151,6 +169,11 @@ data:
       - source_labels: [__meta_kubernetes_service_name]
         action: replace
         target_label: job
+      
+      metric_relabel_configs:
+      - source_labels: [__name__]
+        action: drop
+        regex: etcd_(debugging|disk|request|server).*
 
     # Example scrape config for probing services via the Blackbox Exporter.
     #