Monitor MegaRAID on our storage nodes

Add a MegaCli to Prometheus Text format Exporter for the Storage nodes.

• Add MegaCli
• Add megacli2prom.py script
• Add systemd service (or crontab or so) to run that every five minutes or so
• Add a dashboard and alerts in Grafana

Fixes privatestorageops#19.

So far only a super simple dashboard, no drive errors or so in lieu of experience here, just the number degraded arrays and an alert if that is greater than 0:

morph/grid/production/grid.nix

@@ -70,6 +70,10 @@ let
       # Slightly awkwardly, enable some of our hardware / network / bootloader options.
       ../../../nixos/modules/100tb.nix
 
+      # At least some of our storage nodes utilize MegaRAID storage controllers.
+      # Monitor their array status.
+      ../../../nixos/modules/monitoring/exporters/megacli2prom.nix
+
       # Get all of the configuration that is common across all storage nodes.
       gridlib.storage
 
@@ -88,6 +92,10 @@ let
     # name is quoted because `1` makes `100tb` look an awful lot like a
     # number.
    "100tb".config = nodecfg;
+
+    # Enable statistics gathering for MegaRAID cards.
+    # TODO would be nice to enable only on machines that have such a device.
+    services.private-storage.monitoring.megacli2prom.enable = true;
   };
 
   # Define all of the storage nodes for this grid.