[BUG] Empty Grafana Dashboards

[conradwt]

Describe the bug

After going through the prom_ex.ex file, I'm seeing empty dashboards for a Phoenix application.

Next, I verified that I was using the same Grafana data source name in both the Phoenix app and within Grafana UI.

config/config.exs:

# Configures prom_ex
config :zero_phoenix, ZeroPhoenix.PromEx,
  disabled: false,
  manual_metrics_start_delay: :no_delay,
  drop_metrics_groups: [],
  grafana: [
    host: System.get_env("GRAFANA_CLOUD_URL", "http://localhost:3000"),
    username: System.get_env("GRAFANA_USERNAME", "admin"),
    password: System.get_env("GRAFANA_PASSWORD", "admin"),
    # This is an optional setting and will default to `true`
    folder_name: "zero-to-graphql-using-elixir",
    upload_dashboards_on_start: true
  ]

config :zero_phoenix, ZeroPhoenix.PromEx,
  grafana_datasource_id: System.get_env("GRAFANA_DATASOURCE_ID", "Local Prometheus")

docker-compose.yaml:

version: '3.8'

services:
  grafana:
    # image: grafana/grafana-enterprise
    image: grafana/grafana-oss
    container_name: grafana
    restart: unless-stopped
    environment:
      - GF_SERVER_ROOT_URL=http://localhost:3000
      - GF_FEATURE_TOGGLES_ENABLE=publicDashboards
    #  - GF_INSTALL_PLUGINS=grafana-clock-panel
    ports:
      - '3000:3000'
    volumes:
      - 'grafana_storage:/var/lib/grafana'

  prometheus:
    image: prom/prometheus:v2.46.0
    container_name: prometheus
    ports:
      - 9090:9090
    volumes:
      - ./prometheus:/etc/prometheus/
      - prometheus-data:/prometheus
    command:
      - '--config.file=/etc/prometheus/prometheus.yml'
      - '--storage.tsdb.path=/prometheus'
      - '--web.console.libraries=/etc/prometheus/console_libraries'
      - '--web.console.templates=/etc/prometheus/consoles'
      - '--web.enable-lifecycle'

volumes:
  grafana_storage: {}
  prometheus-data: {}

prom_ex.ex:

defmodule ZeroPhoenix.PromEx do
  use PromEx, otp_app: :zero_phoenix

  alias PromEx.Plugins

  @impl true
  def plugins do
    [
      # PromEx built in plugins
      Plugins.Application,
      Plugins.Beam,
      {Plugins.Phoenix, router: ZeroPhoenixWeb.Router, endpoint: ZeroPhoenixWeb.Endpoint},
      Plugins.Ecto,
      # Plugins.Oban,
      Plugins.PhoenixLiveView,
      Plugins.Absinthe
      # Plugins.Broadway,

      # Add your own PromEx metrics plugins
      # ZeroPhoenix.Users.PromExPlugin
    ]
  end

  @impl true
  def dashboard_assigns do
    [
      # datasource_id: "Local Prometheus",
      datasource_id: Application.get_env(:zero_phoenix, :grafana_datasource_id),
      default_selected_interval: "30s"
    ]
  end

  @impl true
  def dashboards do
    [
      # PromEx built in Grafana dashboards
      {:prom_ex, "application.json"},
      {:prom_ex, "beam.json"},
      {:prom_ex, "phoenix.json"},
      {:prom_ex, "ecto.json"},
      # {:prom_ex, "oban.json"},
      {:prom_ex, "phoenix_live_view.json"},
      {:prom_ex, "absinthe.json"}
      # {:prom_ex, "broadway.json"},

      # Add your dashboard definitions here with the format: {:otp_app, "path_in_priv"}
      # {:zero_phoenix, "/grafana_dashboards/user_metrics.json"}
    ]
  end
end

To Reproduce Steps to reproduce the behavior:

1. Running both Grafana and Prometheus via Docker Compose
2. Dashboards are uploaded to Grafana
3. curl http://localhost:4000/metrics
4. open http://localhost:3000/dashboards

Expected behavior

I would expect to see some data in each one of the dashboards.

Environment

• Elixir version: 1.15.5
• Erlang/OTP version: 26.0.2
• Grafana version:
• Prometheus version: v2.46.0

Additional context

I'm not sure what's the ideal setup for PromEx. The application is being run on http://localhost:4000/graphiql (UI) or http://localhost:4000/graphql (API). The Phoenix app doesn't use Docker and simply runs on macOS 13.5.1 (Ventura). However, I'm using Docker Compose to run both Grafana (http://localhost:3000) and Prometheus (http://localhost:9090) within my local development environment. Finally, I'm using Bandit instead of Cowboy for the webserver.

[conradwt]

My issue was that I was using the wrong URL when using Docker Desktop. Thus, I needed to do the following:

change

  - job_name: 'phoenix_app'
    scrape_interval: 5s
    static_configs:
      - targets: ['localhost:4000']

to

  - job_name: 'phoenix_app'
    scrape_interval: 5s
    static_configs:
      - targets: ['host.docker.internal:4000']

Now, I'm seeing the expected data within all the Grafana dashboards. Thus, I'm going to close this issue.