mcp: move config to .mcp.json.example, gitignore real config

The real .mcp.json now contains Loki credentials for basic auth,
so it should not be committed. The example file has placeholders.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.gitignore

@@ -2,6 +2,9 @@
 result
 result-*
 
+# MCP config (contains secrets)
+.mcp.json
+
 # Terraform/OpenTofu
 terraform/.terraform/
 terraform/.terraform.lock.hcl

.mcp.json.example

@@ -20,7 +20,9 @@
       "env": {
         "PROMETHEUS_URL": "https://prometheus.home.2rjus.net",
         "ALERTMANAGER_URL": "https://alertmanager.home.2rjus.net",
-        "LOKI_URL": "http://monitoring01.home.2rjus.net:3100"
+        "LOKI_URL": "https://loki.home.2rjus.net",
+        "LOKI_USERNAME": "promtail",
+        "LOKI_PASSWORD": "<password from: bao kv get -field=password secret/shared/loki/push-auth>"
       }
     },
     "homelab-deploy": {
@@ -44,4 +46,3 @@
     }
   }
 }
-

docs/plans/monitoring-migration-victoriametrics.md

@@ -14,8 +14,8 @@ a `monitoring` CNAME for seamless transition.
 - Alertmanager (routes to alerttonotify webhook)
 - Grafana (dashboards, datasources)
 - Loki (log aggregation from all hosts via Promtail)
-- Tempo (distributed tracing)
+- Tempo (distributed tracing) - not actively used
-- Pyroscope (continuous profiling)
+- Pyroscope (continuous profiling) - not actively used
 
 **Hardcoded References to monitoring01:**
 - `system/monitoring/logs.nix` - Promtail sends logs to `http://monitoring01.home.2rjus.net:3100`
@@ -44,9 +44,7 @@ If multi-year retention with downsampling becomes necessary later, Thanos can be
                      │  VictoriaMetrics│
                      │  + Grafana      │
      monitoring      │  + Loki         │
-     CNAME ──────────│  + Tempo        │
+     CNAME ──────────│  + Alertmanager │
-                     │  + Pyroscope    │
-                     │  + Alertmanager │
                      │  (vmalert)      │
                      └─────────────────┘
                             ▲
@@ -61,53 +59,48 @@ If multi-year retention with downsampling becomes necessary later, Thanos can be
 
 ## Implementation Plan
 
-### Phase 1: Create monitoring02 Host
+### Phase 1: Create monitoring02 Host [COMPLETE]
 
-Use `create-host` script which handles flake.nix and terraform/vms.tf automatically.
+Host created and deployed at 10.69.13.24 (prod tier) with:
-
+- 4 CPU cores, 8GB RAM, 60GB disk
-1. **Run create-host**: `nix develop -c create-host monitoring02 10.69.13.24`
+- Vault integration enabled
-2. **Update VM resources** in `terraform/vms.tf`:
+- NATS-based remote deployment enabled
-   - 4 cores (same as monitoring01)
+- Grafana with Kanidm OIDC deployed as test instance (`grafana-test.home.2rjus.net`)
-   - 8GB RAM (double, for VictoriaMetrics headroom)
-   - 100GB disk (for 3+ months retention with compression)
-3. **Update host configuration**: Import monitoring services
-4. **Create Vault AppRole**: Add to `terraform/vault/approle.tf`
 
 ### Phase 2: Set Up VictoriaMetrics Stack
 
-Create new service module at `services/monitoring/victoriametrics/` for testing alongside existing
+New service module at `services/victoriametrics/` for VictoriaMetrics + vmalert + Alertmanager.
-Prometheus config. Once validated, this can replace the Prometheus module.
+Imported by monitoring02 alongside the existing Grafana service.
 
-1. **VictoriaMetrics** (port 8428):
+1. **VictoriaMetrics** (port 8428): [DONE]
    - `services.victoriametrics.enable = true`
-   - `services.victoriametrics.retentionPeriod = "3m"` (3 months, increase later based on disk usage)
+   - `retentionPeriod = "3"` (3 months)
-   - Migrate scrape configs via `prometheusConfig`
+   - All scrape configs migrated from Prometheus (22 jobs including auto-generated)
-   - Use native push support (replaces Pushgateway)
+   - Static user override (DynamicUser disabled) for credential file access
+   - OpenBao token fetch service + 30min refresh timer
+   - Apiary bearer token via vault.secrets
 
-2. **vmalert** for alerting rules:
+2. **vmalert** for alerting rules: [DONE]
-   - `services.vmalert.enable = true`
+   - Points to VictoriaMetrics datasource at localhost:8428
-   - Point to VictoriaMetrics for metrics evaluation
+   - Reuses existing `services/monitoring/rules.yml` directly via `settings.rule`
-   - Keep rules in separate `rules.yml` file (same format as Prometheus)
+   - No notifier configured during parallel operation (prevents duplicate alerts)
-   - No receiver configured during parallel operation (prevents duplicate alerts)
 
-3. **Alertmanager** (port 9093):
+3. **Alertmanager** (port 9093): [DONE]
-   - Keep existing configuration (alerttonotify webhook routing)
+   - Same configuration as monitoring01 (alerttonotify webhook routing)
-   - Only enable receiver after cutover from monitoring01
+   - Will only receive alerts after cutover (vmalert notifier disabled)
 
-4. **Loki** (port 3100):
+4. **Grafana** (port 3000): [DONE]
-   - Same configuration as current
+   - VictoriaMetrics datasource (localhost:8428) as default
+   - monitoring01 Prometheus datasource kept for comparison during parallel operation
+   - Loki datasource pointing to localhost (after Loki migrated to monitoring02)
 
-5. **Grafana** (port 3000):
+5. **Loki** (port 3100): [DONE]
-   - Define dashboards declaratively via NixOS options (not imported from monitoring01)
+   - Same configuration as monitoring01 in standalone `services/loki/` module
-   - Reference existing dashboards on monitoring01 for content inspiration
+   - Grafana datasource updated to localhost:3100
-   - Configure VictoriaMetrics datasource (port 8428)
-   - Configure Loki datasource
 
-6. **Tempo** (ports 3200, 3201):
+**Note:** pve-exporter and pushgateway scrape targets are not included on monitoring02.
-   - Same configuration
+pve-exporter requires a local exporter instance; pushgateway is replaced by VictoriaMetrics
-
+native push support.
-7. **Pyroscope** (port 4040):
-   - Same Docker-based deployment
 
 ### Phase 3: Parallel Operation
 
@@ -147,7 +140,6 @@ Update hardcoded references to use the CNAME:
    - prometheus.home.2rjus.net -> monitoring.home.2rjus.net:8428
    - alertmanager.home.2rjus.net -> monitoring.home.2rjus.net:9093
    - grafana.home.2rjus.net -> monitoring.home.2rjus.net:3000
-   - pyroscope.home.2rjus.net -> monitoring.home.2rjus.net:4040
 
 Note: `hosts/template2/bootstrap.nix` stays pointed at monitoring01 until decommission.
 
@@ -171,24 +163,9 @@ Once ready to cut over:
 
 ## Current Progress
 
-### monitoring02 Host Created (2026-02-08)
+- **Phase 1** complete (2026-02-08): monitoring02 host created, Grafana with Kanidm OIDC validated
-
+- **Phase 2** complete (2026-02-17): VictoriaMetrics, vmalert, Alertmanager, Loki, Grafana datasources configured
-Host deployed at 10.69.13.24 (test tier) with:
+  - Tempo and Pyroscope deferred (not actively used; can be added later if needed)
-- 4 CPU cores, 8GB RAM, 60GB disk
-- Vault integration enabled
-- NATS-based remote deployment enabled
-
-### Grafana with Kanidm OIDC (2026-02-08)
-
-Grafana deployed on monitoring02 as a test instance (`grafana-test.home.2rjus.net`):
-- Kanidm OIDC authentication (PKCE enabled)
-- Role mapping: `admins` → Admin, others → Viewer
-- Declarative datasources pointing to monitoring01 (Prometheus, Loki)
-- Local Caddy for TLS termination via internal ACME CA
-
-This validates the Grafana + OIDC pattern before the full VictoriaMetrics migration. The existing
-`services/monitoring/grafana.nix` on monitoring01 can be replaced with the new `services/grafana/`
-module once monitoring02 becomes the primary monitoring host.
 
 ## Open Questions
 
@@ -198,31 +175,14 @@ module once monitoring02 becomes the primary monitoring host.
 
 ## VictoriaMetrics Service Configuration
 
-Example NixOS configuration for monitoring02:
+Implemented in `services/victoriametrics/default.nix`. Key design decisions:
 
-```nix
+- **Static user**: VictoriaMetrics NixOS module uses `DynamicUser`, overridden with a static
-# VictoriaMetrics replaces Prometheus
+  `victoriametrics` user so vault.secrets and credential files work correctly
-services.victoriametrics = {
+- **Shared rules**: vmalert reuses `services/monitoring/rules.yml` via `settings.rule` path
-  enable = true;
+  reference (no YAML-to-Nix conversion needed)
-  retentionPeriod = "3m";  # 3 months, increase based on disk usage
+- **Scrape config reuse**: Uses the same `lib/monitoring.nix` functions and
-  prometheusConfig = {
+  `services/monitoring/external-targets.nix` as Prometheus for auto-generated targets
-    global.scrape_interval = "15s";
-    scrape_configs = [
-      # Auto-generated node-exporter targets
-      # Service-specific scrape targets
-      # External targets
-    ];
-  };
-};
-
-# vmalert for alerting rules (no receiver during parallel operation)
-services.vmalert = {
-  enable = true;
-  datasource.url = "http://localhost:8428";
-  # notifier.alertmanager.url = "http://localhost:9093";  # Enable after cutover
-  rule = [ ./rules.yml ];
-};
-```
 
 ## Rollback Plan
 

hosts/monitoring02/configuration.nix

@@ -18,8 +18,7 @@
     role = "monitoring";
   };
 
-  # DNS CNAME for Grafana test instance
+  homelab.dns.cnames = [ "grafana-test" "metrics" "vmalert" "loki" ];
-  homelab.dns.cnames = [ "grafana-test" ];
 
   # Enable Vault integration
   vault.enable = true;

hosts/monitoring02/default.nix

@@ -2,5 +2,7 @@
   imports = [
     ./configuration.nix
     ../../services/grafana
+    ../../services/victoriametrics
+    ../../services/loki
   ];
 }

services/grafana/default.nix

@@ -34,21 +34,27 @@
       };
     };
 
-    # Declarative datasources pointing to monitoring01
+    # Declarative datasources
     provision.datasources.settings = {
       apiVersion = 1;
       datasources = [
         {
-          name = "Prometheus";
+          name = "VictoriaMetrics";
+          type = "prometheus";
+          url = "http://localhost:8428";
+          isDefault = true;
+          uid = "victoriametrics";
+        }
+        {
+          name = "Prometheus (monitoring01)";
           type = "prometheus";
           url = "http://monitoring01.home.2rjus.net:9090";
-          isDefault = true;
           uid = "prometheus";
         }
         {
           name = "Loki";
           type = "loki";
-          url = "http://monitoring01.home.2rjus.net:3100";
+          url = "http://localhost:3100";
           uid = "loki";
         }
       ];
@@ -81,22 +87,20 @@
   services.caddy = {
     enable = true;
     package = pkgs.unstable.caddy;
-    configFile = pkgs.writeText "Caddyfile" ''
+    globalConfig = ''
-      {
       acme_ca https://vault.home.2rjus.net:8200/v1/pki_int/acme/directory
       metrics
-      }
+    '';
-
+    virtualHosts."grafana-test.home.2rjus.net".extraConfig = ''
-      grafana-test.home.2rjus.net {
       log {
         output file /var/log/caddy/grafana.log {
           mode 644
         }
       }
-
       reverse_proxy http://127.0.0.1:3000
-      }
+    '';
-
+    # Metrics endpoint on plain HTTP for Prometheus scraping
+    extraConfig = ''
       http://${config.networking.hostName}.home.2rjus.net/metrics {
         metrics
       }

services/loki/default.nix

@@ -0,0 +1,104 @@
+{ config, lib, pkgs, ... }:
+let
+  # Script to generate bcrypt hash from Vault password for Caddy basic_auth
+  generateCaddyAuth = pkgs.writeShellApplication {
+    name = "generate-caddy-loki-auth";
+    runtimeInputs = [ config.services.caddy.package ];
+    text = ''
+      PASSWORD=$(cat /run/secrets/loki-push-auth)
+      HASH=$(caddy hash-password --plaintext "$PASSWORD")
+      echo "LOKI_PUSH_HASH=$HASH" > /run/secrets/caddy-loki-auth.env
+      chmod 0400 /run/secrets/caddy-loki-auth.env
+    '';
+  };
+in
+{
+  # Fetch Loki push password from Vault
+  vault.secrets.loki-push-auth = {
+    secretPath = "shared/loki/push-auth";
+    extractKey = "password";
+    services = [ "caddy" ];
+  };
+
+  # Generate bcrypt hash for Caddy before it starts
+  systemd.services.caddy-loki-auth = {
+    description = "Generate Caddy basic auth hash for Loki";
+    after = [ "vault-secret-loki-push-auth.service" ];
+    requires = [ "vault-secret-loki-push-auth.service" ];
+    before = [ "caddy.service" ];
+    requiredBy = [ "caddy.service" ];
+    serviceConfig = {
+      Type = "oneshot";
+      RemainAfterExit = true;
+      ExecStart = lib.getExe generateCaddyAuth;
+    };
+  };
+
+  # Load the bcrypt hash as environment variable for Caddy
+  services.caddy.environmentFile = "/run/secrets/caddy-loki-auth.env";
+
+  # Caddy reverse proxy for Loki with basic auth
+  services.caddy.virtualHosts."loki.home.2rjus.net".extraConfig = ''
+    basic_auth {
+      promtail {env.LOKI_PUSH_HASH}
+    }
+    reverse_proxy http://127.0.0.1:3100
+  '';
+
+  services.loki = {
+    enable = true;
+    configuration = {
+      auth_enabled = false;
+
+      server = {
+        http_listen_address = "127.0.0.1";
+        http_listen_port = 3100;
+      };
+      common = {
+        ring = {
+          instance_addr = "127.0.0.1";
+          kvstore = {
+            store = "inmemory";
+          };
+        };
+        replication_factor = 1;
+        path_prefix = "/var/lib/loki";
+      };
+      schema_config = {
+        configs = [
+          {
+            from = "2024-01-01";
+            store = "tsdb";
+            object_store = "filesystem";
+            schema = "v13";
+            index = {
+              prefix = "loki_index_";
+              period = "24h";
+            };
+          }
+        ];
+      };
+      storage_config = {
+        filesystem = {
+          directory = "/var/lib/loki/chunks";
+        };
+      };
+      compactor = {
+        working_directory = "/var/lib/loki/compactor";
+        compaction_interval = "10m";
+        retention_enabled = true;
+        retention_delete_delay = "2h";
+        retention_delete_worker_count = 150;
+        delete_request_store = "filesystem";
+      };
+      limits_config = {
+        retention_period = "30d";
+        ingestion_rate_mb = 10;
+        ingestion_burst_size_mb = 20;
+        max_streams_per_user = 10000;
+        max_query_series = 500;
+        max_query_parallelism = 8;
+      };
+    };
+  };
+}

services/victoriametrics/default.nix

@@ -0,0 +1,219 @@
+{ self, config, lib, pkgs, ... }:
+let
+  monLib = import ../../lib/monitoring.nix { inherit lib; };
+  externalTargets = import ../monitoring/external-targets.nix;
+
+  nodeExporterTargets = monLib.generateNodeExporterTargets self externalTargets;
+  autoScrapeConfigs = monLib.generateScrapeConfigs self externalTargets;
+
+  # Script to fetch AppRole token for VictoriaMetrics to use when scraping OpenBao metrics
+  fetchOpenbaoToken = pkgs.writeShellApplication {
+    name = "fetch-openbao-token-vm";
+    runtimeInputs = [ pkgs.curl pkgs.jq ];
+    text = ''
+      VAULT_ADDR="https://vault01.home.2rjus.net:8200"
+      APPROLE_DIR="/var/lib/vault/approle"
+      OUTPUT_FILE="/run/secrets/victoriametrics/openbao-token"
+
+      # Read AppRole credentials
+      if [ ! -f "$APPROLE_DIR/role-id" ] || [ ! -f "$APPROLE_DIR/secret-id" ]; then
+        echo "AppRole credentials not found at $APPROLE_DIR" >&2
+        exit 1
+      fi
+
+      ROLE_ID=$(cat "$APPROLE_DIR/role-id")
+      SECRET_ID=$(cat "$APPROLE_DIR/secret-id")
+
+      # Authenticate to Vault
+      AUTH_RESPONSE=$(curl -sf -k -X POST \
+        -d "{\"role_id\":\"$ROLE_ID\",\"secret_id\":\"$SECRET_ID\"}" \
+        "$VAULT_ADDR/v1/auth/approle/login")
+
+      # Extract token
+      VAULT_TOKEN=$(echo "$AUTH_RESPONSE" | jq -r '.auth.client_token')
+      if [ -z "$VAULT_TOKEN" ] || [ "$VAULT_TOKEN" = "null" ]; then
+        echo "Failed to extract Vault token from response" >&2
+        exit 1
+      fi
+
+      # Write token to file
+      mkdir -p "$(dirname "$OUTPUT_FILE")"
+      echo -n "$VAULT_TOKEN" > "$OUTPUT_FILE"
+      chown victoriametrics:victoriametrics "$OUTPUT_FILE"
+      chmod 0400 "$OUTPUT_FILE"
+
+      echo "Successfully fetched OpenBao token"
+    '';
+  };
+
+  scrapeConfigs = [
+    # Auto-generated node-exporter targets from flake hosts + external
+    {
+      job_name = "node-exporter";
+      static_configs = nodeExporterTargets;
+    }
+    # Systemd exporter on all hosts (same targets, different port)
+    {
+      job_name = "systemd-exporter";
+      static_configs = map
+        (cfg: cfg // {
+          targets = map (t: builtins.replaceStrings [ ":9100" ] [ ":9558" ] t) cfg.targets;
+        })
+        nodeExporterTargets;
+    }
+    # Local monitoring services
+    {
+      job_name = "victoriametrics";
+      static_configs = [{ targets = [ "localhost:8428" ]; }];
+    }
+    {
+      job_name = "loki";
+      static_configs = [{ targets = [ "localhost:3100" ]; }];
+    }
+    {
+      job_name = "grafana";
+      static_configs = [{ targets = [ "localhost:3000" ]; }];
+    }
+    {
+      job_name = "alertmanager";
+      static_configs = [{ targets = [ "localhost:9093" ]; }];
+    }
+    # Caddy metrics from nix-cache02
+    {
+      job_name = "nix-cache_caddy";
+      scheme = "https";
+      static_configs = [{ targets = [ "nix-cache.home.2rjus.net" ]; }];
+    }
+    # OpenBao metrics with bearer token auth
+    {
+      job_name = "openbao";
+      scheme = "https";
+      metrics_path = "/v1/sys/metrics";
+      params = { format = [ "prometheus" ]; };
+      static_configs = [{ targets = [ "vault01.home.2rjus.net:8200" ]; }];
+      authorization = {
+        type = "Bearer";
+        credentials_file = "/run/secrets/victoriametrics/openbao-token";
+      };
+    }
+    # Apiary external service
+    {
+      job_name = "apiary";
+      scheme = "https";
+      scrape_interval = "60s";
+      static_configs = [{ targets = [ "apiary.t-juice.club" ]; }];
+      authorization = {
+        type = "Bearer";
+        credentials_file = "/run/secrets/victoriametrics-apiary-token";
+      };
+    }
+  ] ++ autoScrapeConfigs;
+in
+{
+  # Static user for VictoriaMetrics (overrides DynamicUser) so vault.secrets
+  # and credential files can be owned by this user
+  users.users.victoriametrics = {
+    isSystemUser = true;
+    group = "victoriametrics";
+  };
+  users.groups.victoriametrics = { };
+
+  # Override DynamicUser since we need a static user for credential file access
+  systemd.services.victoriametrics.serviceConfig = {
+    DynamicUser = lib.mkForce false;
+    User = "victoriametrics";
+    Group = "victoriametrics";
+  };
+
+  # Systemd service to fetch AppRole token for OpenBao scraping
+  systemd.services.victoriametrics-openbao-token = {
+    description = "Fetch OpenBao token for VictoriaMetrics metrics scraping";
+    after = [ "network-online.target" ];
+    wants = [ "network-online.target" ];
+    before = [ "victoriametrics.service" ];
+    requiredBy = [ "victoriametrics.service" ];
+
+    serviceConfig = {
+      Type = "oneshot";
+      ExecStart = lib.getExe fetchOpenbaoToken;
+    };
+  };
+
+  # Timer to periodically refresh the token (AppRole tokens have 1-hour TTL)
+  systemd.timers.victoriametrics-openbao-token = {
+    description = "Refresh OpenBao token for VictoriaMetrics";
+    wantedBy = [ "timers.target" ];
+    timerConfig = {
+      OnBootSec = "5min";
+      OnUnitActiveSec = "30min";
+      RandomizedDelaySec = "5min";
+    };
+  };
+
+  # Fetch apiary bearer token from Vault
+  vault.secrets.victoriametrics-apiary-token = {
+    secretPath = "hosts/monitoring01/apiary-token";
+    extractKey = "password";
+    owner = "victoriametrics";
+    group = "victoriametrics";
+    services = [ "victoriametrics" ];
+  };
+
+  services.victoriametrics = {
+    enable = true;
+    retentionPeriod = "3"; # 3 months
+    # Disable config check since we reference external credential files
+    checkConfig = false;
+    prometheusConfig = {
+      global.scrape_interval = "15s";
+      scrape_configs = scrapeConfigs;
+    };
+  };
+
+  # vmalert for alerting rules - no notifier during parallel operation
+  services.vmalert.instances.default = {
+    enable = true;
+    settings = {
+      "datasource.url" = "http://localhost:8428";
+      # Blackhole notifications during parallel operation to prevent duplicate alerts.
+      # Replace with notifier.url after cutover from monitoring01:
+      # "notifier.url" = [ "http://localhost:9093" ];
+      "notifier.blackhole" = true;
+      "rule" = [ ../monitoring/rules.yml ];
+    };
+  };
+
+  # Caddy reverse proxy for VictoriaMetrics and vmalert
+  services.caddy.virtualHosts."metrics.home.2rjus.net".extraConfig = ''
+    reverse_proxy http://127.0.0.1:8428
+  '';
+  services.caddy.virtualHosts."vmalert.home.2rjus.net".extraConfig = ''
+    reverse_proxy http://127.0.0.1:8880
+  '';
+
+  # Alertmanager - same config as monitoring01 but will only receive
+  # alerts after cutover (vmalert notifier is disabled above)
+  services.prometheus.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";
+            }
+          ];
+        }
+      ];
+    };
+  };
+}

system/monitoring/logs.nix

@@ -16,6 +16,16 @@ in
       SystemKeepFree=1G
     '';
   };
+
+  # Fetch Loki push password from Vault (only on hosts with Vault enabled)
+  vault.secrets.promtail-loki-auth = lib.mkIf config.vault.enable {
+    secretPath = "shared/loki/push-auth";
+    extractKey = "password";
+    owner = "promtail";
+    group = "promtail";
+    services = [ "promtail" ];
+  };
+
   # Configure promtail
   services.promtail = {
     enable = true;
@@ -31,6 +41,14 @@ in
         {
           url = "http://monitoring01.home.2rjus.net:3100/loki/api/v1/push";
         }
+      ] ++ lib.optionals config.vault.enable [
+        {
+          url = "https://loki.home.2rjus.net/loki/api/v1/push";
+          basic_auth = {
+            username = "promtail";
+            password_file = "/run/secrets/promtail-loki-auth";
+          };
+        }
       ];
 
       scrape_configs = [

terraform/vault/approle.tf

@@ -26,6 +26,17 @@ path "secret/data/shared/nixos-exporter/*" {
 EOT
 }
 
+# Shared policy for Loki push authentication (all hosts push logs)
+resource "vault_policy" "loki_push" {
+  name = "loki-push"
+
+  policy = <<EOT
+path "secret/data/shared/loki/*" {
+  capabilities = ["read", "list"]
+}
+EOT
+}
+
 # Define host access policies
 locals {
   host_policies = {
@@ -78,7 +89,7 @@ locals {
       ]
     }
 
-    # Wave 3: DNS servers
+    # Wave 3: DNS servers (managed in hosts-generated.tf)
 
     # Wave 4: http-proxy
     "http-proxy" = {
@@ -104,10 +115,11 @@ locals {
       ]
     }
 
-    # monitoring02: Grafana test instance
+    # monitoring02: Grafana + VictoriaMetrics
     "monitoring02" = {
       paths = [
         "secret/data/hosts/monitoring02/*",
+        "secret/data/hosts/monitoring01/apiary-token",
         "secret/data/services/grafana/*",
       ]
     }
@@ -137,7 +149,7 @@ resource "vault_approle_auth_backend_role" "hosts" {
   backend   = vault_auth_backend.approle.path
   role_name = each.key
   token_policies = concat(
-    ["${each.key}-policy", "homelab-deploy", "nixos-exporter"],
+    ["${each.key}-policy", "homelab-deploy", "nixos-exporter", "loki-push"],
     lookup(each.value, "extra_policies", [])
   )
 

terraform/vault/hosts-generated.tf

@@ -74,7 +74,7 @@ resource "vault_approle_auth_backend_role" "generated_hosts" {
 
   backend            = vault_auth_backend.approle.path
   role_name          = each.key
-  token_policies     = ["host-${each.key}", "homelab-deploy", "nixos-exporter"]
+  token_policies     = ["host-${each.key}", "homelab-deploy", "nixos-exporter", "loki-push"]
   secret_id_ttl      = 0 # Never expire (wrapped tokens provide time limit)
   token_ttl          = 3600
   token_max_ttl      = 3600

terraform/vault/secrets.tf

@@ -153,6 +153,12 @@ locals {
       auto_generate   = true
       password_length = 64
     }
+
+    # Loki push authentication (used by Promtail on all hosts)
+    "shared/loki/push-auth" = {
+      auto_generate   = true
+      password_length = 32
+    }
   }
 }