Add script to monitor GPFS

Reviewed-by: Aleix Boné <abonerib@bsc.es>

m/hut/gpfs-probe.nix

@@ -0,0 +1,31 @@
+{ pkgs, config, lib, ... }:
+let
+  gpfs-probe-script = pkgs.runCommand "gpfs-probe.sh" { }
+    ''
+      cp ${./gpfs-probe.sh} $out;
+      chmod +x $out
+    ''
+  ;
+in
+{
+  # Use a new user to handle the SSH keys
+  users.groups.ssh-robot = { };
+  users.users.ssh-robot = {
+    description = "SSH Robot";
+    isNormalUser = true;
+    home = "/var/lib/ssh-robot";
+  };
+
+  systemd.services.gpfs-probe = {
+    description = "Daemon to report GPFS latency via SSH";
+    path = [ pkgs.openssh pkgs.netcat ];
+    after = [ "network.target" ];
+    wantedBy = [ "default.target" ];
+    serviceConfig = {
+      Type = "simple";
+      ExecStart = "${pkgs.socat}/bin/socat TCP4-LISTEN:9966,fork EXEC:${gpfs-probe-script}";
+      User = "ssh-robot";
+      Group = "ssh-robot";
+    };
+  };
+}

m/hut/gpfs-probe.sh

@@ -0,0 +1,18 @@
+#!/bin/sh
+
+N=500
+
+t=$(timeout 5 ssh bsc015557@glogin2.bsc.es "timeout 3 command time -f %e touch /gpfs/projects/bsc15/bsc015557/gpfs.{1..$N} 2>&1; rm -f /gpfs/projects/bsc15/bsc015557/gpfs.{1..$N}")
+
+if [ -z "$t" ]; then
+  t="5.00"
+fi
+
+cat <<EOF
+HTTP/1.1 200 OK
+Content-Type: text/plain; version=0.0.4; charset=utf-8; escaping=values
+
+# HELP gpfs_touch_latency Time to create $N files.
+# TYPE gpfs_touch_latency gauge
+gpfs_touch_latency $t
+EOF

m/hut/monitoring.nix

@@ -1,7 +1,10 @@
 { config, lib, ... }:
 
 {
-  imports = [ ../module/slurm-exporter.nix ];
+  imports = [
+    ../module/slurm-exporter.nix
+    ./gpfs-probe.nix
+  ];
 
   age.secrets.grafanaJungleRobotPassword = {
     file = ../../secrets/jungle-robot-password.age;