nixos-servers/services/monitoring/prometheus.nix

{ self, lib, ... }:
let
  monLib = import ../../lib/monitoring.nix { inherit lib; };
  externalTargets = import ./external-targets.nix;

  nodeExporterTargets = monLib.generateNodeExporterTargets self externalTargets;
  autoScrapeConfigs = monLib.generateScrapeConfigs self externalTargets;
in
{
  services.prometheus = {
    enable = true;
    alertmanager = {
      enable = true;
      configuration = {
        global = {
        };
        route = {
          receiver = "webhook_natstonotify";
          group_wait = "30s";
          group_interval = "5m";
          repeat_interval = "1h";
          group_by = [ "alertname" ];
        };
        receivers = [
          {
            name = "webhook_natstonotify";
            webhook_configs = [
              {
                url = "http://localhost:5001/alert";
              }
            ];
          }
        ];
      };
    };
    alertmanagers = [
      {
        static_configs = [
          {
            targets = [ "localhost:9093" ];
          }
        ];
      }
    ];

    retentionTime = "30d";
    globalConfig = {
      scrape_interval = "15s";
    };
    rules = [
      (builtins.readFile ./rules.yml)
    ];

    scrapeConfigs = [
      # Auto-generated node-exporter targets from flake hosts + external
      {
        job_name = "node-exporter";
        static_configs = [
          {
            targets = nodeExporterTargets;
          }
        ];
      }
      # Local monitoring services (not auto-generated)
      {
        job_name = "prometheus";
        static_configs = [
          {
            targets = [ "localhost:9090" ];
          }
        ];
      }
      {
        job_name = "loki";
        static_configs = [
          {
            targets = [ "localhost:3100" ];
          }
        ];
      }
      {
        job_name = "grafana";
        static_configs = [
          {
            targets = [ "localhost:3000" ];
          }
        ];
      }
      {
        job_name = "alertmanager";
        static_configs = [
          {
            targets = [ "localhost:9093" ];
          }
        ];
      }
      {
        job_name = "pushgateway";
        honor_labels = true;
        static_configs = [
          {
            targets = [ "localhost:9091" ];
          }
        ];
      }
      {
        job_name = "labmon";
        static_configs = [
          {
            targets = [ "monitoring01.home.2rjus.net:9969" ];
          }
        ];
      }
      # TODO: nix-cache_caddy can't be auto-generated because the cert is issued
      # for nix-cache.home.2rjus.net (service CNAME), not nix-cache01 (hostname).
      # Consider adding a target override to homelab.monitoring.scrapeTargets.
      {
        job_name = "nix-cache_caddy";
        scheme = "https";
        static_configs = [
          {
            targets = [ "nix-cache.home.2rjus.net" ];
          }
        ];
      }
      # pve-exporter with complex relabel config
      {
        job_name = "pve-exporter";
        static_configs = [
          {
            targets = [ "10.69.12.75" ];
          }
        ];
        metrics_path = "/pve";
        params = {
          module = [ "default" ];
          cluster = [ "1" ];
          node = [ "1" ];
        };
        relabel_configs = [
          {
            source_labels = [ "__address__" ];
            target_label = "__param_target";
          }
          {
            source_labels = [ "__param_target" ];
            target_label = "instance";
          }
          {
            target_label = "__address__";
            replacement = "127.0.0.1:9221";
          }
        ];
      }
    ] ++ autoScrapeConfigs;

    pushgateway = {
      enable = true;
      web = {
        external-url = "https://pushgw.home.2rjus.net";
      };
    };
  };
}