Proper monitoring is coming to the infrastructure underneath all of this, running on Zabbix. Here’s what it is, what it isn’t, and why you’re reading about something most of you will never log into.
What it’s for
NCSGSS — the status page — answers one question: is this service responding right now. That’s the right question when something has already broken, and it’s deliberately simple.
NCSNPM answers the questions underneath that. Disk filling up on a host. Memory pressure building over three days. A network link flapping. A backup that ran but took four times longer than usual. Certificates approaching expiry. A power supply reporting a fault in a machine nobody has looked at in a month.
In other words: the status page tells you the house is on fire. This is the thing that notices the wiring getting warm.
What it means for you
Mostly, less drama.
The majority of outages people actually experience are not sudden. A disk fills over weeks. A service leaks memory until it falls over on a Sunday. A certificate quietly expires on a date somebody wrote down eleven months ago and then forgot. Every one of those is visible in advance if something is looking.
So the practical effect is that more problems get caught before they become your problem, and the ones that do reach you get diagnosed faster because there’s history to look at rather than guesswork.
You don’t have to do anything. There’s nothing to install, nothing to sign up for, and no change to how you use any service.
For site admins
If you’re one of the people who helps run this rather than just uses it:
Logging in. NCSNPM sits behind NCSMFA, like everything else. Same credentials, same second factor, no separate account to create.
Alerts. Site admins get added to the NCSNPM group in Nextcloud Talk, and alerts land there.
Talk was chosen over email on purpose: an alert that arrives in a group chat gets seen and can be answered in the same place, rather than sitting in an inbox next to four hundred other things while everyone assumes somebody else is handling it.
One honest limitation
The alerting path runs over the same infrastructure it’s watching.
If a site loses power or its internet connection, NCSNPM can’t tell anyone about it through Nextcloud Talk, because Nextcloud Talk is on the other side of the thing that just broke. The same is true of the status page. In a genuine full outage, the systems designed to announce the outage go down with it.
That’s a known gap rather than an oversight. An out-of-band path — something that doesn’t depend on the site being reachable — is the obvious fix and it’s on the list. Until it exists, a total site failure is still going to be noticed the old-fashioned way: by somebody trying to use something.
When
Coming. Not dated.
The monitoring itself is the easy part; deciding what deserves to wake somebody up is the part that takes time. A monitoring system that cries wolf is worse than none at all, because people learn to ignore it and then ignore the one that mattered. Better to tune it quietly first and announce it when the alerts mean something.
Problems
Open a ticket at ncshds.northlandcentral.org, email support@ncsmx.org, or call (210) 904-8005.
And the usual reminder: if something looks broken and the status page says otherwise, report it anyway. No amount of monitoring sees everything, and the gaps are exactly where the interesting failures live.
