wonderwall/hack/dashboard.yaml

title: Wonderwall
editable: true
tags: [generated, yaml]
auto_refresh: 1m
time: ["now-24h", "now"]
timezone: default # valid values are: utc, browser, default

# Render to JSON using https://github.com/K-Phoen/grabana v0.17.0 or newer
# Import into Grafana using UI (remember to select folder)

variables:
  - custom:
      name: env
      default: dev
      values_map:
        dev: dev
        prod: prod
  - datasource:
      name: ds
      type: prometheus
      regex: $env-gcp
      include_all: true
      hide: variable
  - query:
      name: redis_op
      label: Redis Operation
      datasource: $env-gcp
      request: "label_values(wonderwall_redis_latency_bucket, operation)"
      include_all: true
      default_all: true
      hide: variable

rows:
  - name: Versions
    collapse: false
    panels:
      - single_stat:
          title: Sidecar versions in use
          datasource: $ds
          transparent: true
          targets:
            - prometheus:
                query: count(label_replace(kube_pod_container_info{container="wonderwall"}, "version", "$1", "image", ".*:(.*)")) by (version)
                legend: "{{ version }}"
                instant: true
  - name: Resource usage
    collapse: false
    panels:
      - graph:
          title: Memory usage - $ds
          datasource: $ds
          transparent: true
          span: 4
          targets:
            - prometheus:
                query: avg(kube_pod_container_resource_limits{container="wonderwall",resource="memory"}) by (namespace)
                legend: "limits in {{ namespace }}"
            - prometheus:
                query: avg(kube_pod_container_resource_requests{container="wonderwall",resource="memory"}) by (namespace)
                legend: "requests in {{ namespace }}"
            - prometheus:
                query: sum(container_memory_working_set_bytes{container="wonderwall"}) by (pod, namespace)
                legend: "working set {{ pod }} in {{ namespace }}"
            - prometheus:
                query: sum(container_memory_usage_bytes{container="wonderwall"}) by (pod, namespace)
                legend: "Resident set size {{ pod }} in {{ namespace }}"
      - graph:
          title: CPU usage - $ds
          datasource: $ds
          transparent: true
          span: 4
          targets:
            - prometheus:
                query: avg(kube_pod_container_resource_limits{container="wonderwall",resource="cpu"}) by (namespace)
                legend: "limits in {{ namespace }}"
            - prometheus:
                query: avg(kube_pod_container_resource_requests{container="wonderwall",resource="cpu"}) by (namespace)
                legend: "requests in {{ namespace }}"
            - prometheus:
                query: sum(irate(container_cpu_usage_seconds_total{container="wonderwall"}[2m])) by (pod, namespace)
                legend: "{{ pod }} in {{ namespace }}"
  - name: Redis Latency - $redis_op
    repeat_for: redis_op
    collapse: true
    panels:
      - heatmap:
          # Must be done manually in Grafana after import: Set max datapoints to 25
          title: $ds
          datasource: $ds
          repeat: ds
          data_format: time_series_buckets
          hide_zero_buckets: true
          transparent: true
          span: 4
          tooltip:
            show: true
            showhistogram: false
            decimals: 0
          yaxis:
            unit: "dtdurations"
            decimals: 0
          targets:
            - prometheus:
                query: sum(increase(wonderwall_redis_latency_bucket{operation="$redis_op"}[$__interval])) by (le)
                legend: "{{ le }}"
                format: heatmap