Parallelize CVMFS monitoring to bring back the CVMFS Grafana dashboard

The script `/usr/bin/check_cvmfs_repos` installed by the CVMFS monitoring role `hxr.monitor-cvmfs` takes longer than 2 minutes to run (the Telegraf timeout for this script) due to misbehaving CVMFS servers and serial execution. This results in no measurements being registered.

```
Jul 25 13:52:00 cvmfs1-ufr0.internal.galaxyproject.eu telegraf[2616631]: 2024-07-25T11:52:00Z E! [inputs.exec] Error in plugin: exec: command timed out for command "/usr/bin/check_cvmfs_repos": /usr/bin/check_cvmfs_repos: line 9: [: : integer expression expected...
```

Add timeout to `curl` calls in `check_cvmfs_repos` script from CVMFS monitoring role `hxr.monitor-cvmfs` and parallelize all `check_repo` calls so that the script is guaranteed to exit before it times out.

roles/hxr.monitor-cvmfs/templates/main.sh

@@ -3,8 +3,8 @@ check_repo() {
 	host="$1"
 	repo="$2"
 
-	http_code="$(curl http://$host/cvmfs/$repo/.cvmfspublished -I --silent | head -n 1 | cut -f2 -d' ')"
-	header="$(curl http://$host/cvmfs/$repo/.cvmfspublished --silent | head -n 12)"
+	http_code="$(curl --max-time 20 http://$host/cvmfs/$repo/.cvmfspublished -I --silent | head -n 1 | cut -f2 -d' ')"
+	header="$(curl --max-time 20 http://$host/cvmfs/$repo/.cvmfspublished --silent | head -n 12)"
 
 	if [ "$http_code" -eq "200" ]; then
 		# https://cvmfs.readthedocs.io/en/stable/cpt-details.html#repository-manifest-cvmfspublished
@@ -21,6 +21,8 @@ check_repo() {
 
 {% for host in cvmfs_check_servers.hosts %}
 {% for repo in cvmfs_check_servers.repos %}
-check_repo {{ host }} {{ repo }}
+check_repo {{ host }} {{ repo }} &
 {% endfor %}
 {% endfor %}
+
+wait