A Silent Sensor Is Not a Quiet Network: How to Read an Empty Alert Queue
There is a failure in security monitoring that does not look like a failure. The alert queue is empty. Nothing is red. The shift lead glances at it, sees nothing to do, and moves on. Then it turns out the sensor stopped reporting on Wednesday.
An empty queue and a broken sensor look identical, and only one of them is good news. I built EmberSecurity so the screen does not let you confuse the two.
Two ways to be quiet
A plant can be quiet because nothing is happening. Or it can be quiet because nothing is being watched. Any monitoring tool that only counts alerts will show you the same empty list in both cases.
EmberSecurity runs four containers on every EmberNet edge box, each with one job: a network watch for the PLC leg and the plant network, an OT protocol sensor for the industrial protocols on the wire, a runtime guard for the box itself, and an integrity and posture container for files and configuration. Each one reports on its own.
The first three send a heartbeat, so a container that has stopped is shown as stopped. A silent sensor is never mistaken for a quiet network.
The Command view starts with the right question
The first view in the EmberSecurity area of EmberCORE is Command, and its first question is whether anything is wrong and where. It counts alerts by severity over the last hour, day, or week. Next to that it shows how many boxes are reporting and how many have been silent for more than 15 minutes.
Below that sits a heat strip with one cell per box. Each cell is quiet, alerting, critical, or silent. Silent is its own state, with its own place on the screen, and it is not folded into quiet. A chart of alerts per hour sits beside it, and a Live switch keeps the whole view updating.
The queue checks itself
The same idea runs through every view. An empty queue only means good news while the boxes are reporting, so the queue offers a link to check them. If no box has reported, the views say so instead of drawing an empty chart. If EmberSecurity cannot be reached, the view shows an error, not an empty table.
I care about that last one more than it might sound. An error you can see is a problem you can fix. A blank table that is blank because the data never arrived is a problem that waits for the worst possible day to show itself.
Fireball's own staff get one more safeguard. They see a banner when EmberSecurity's central side needs attention, so an alert list that may be out of date is never mistaken for a calm plant.
One box at a time
When a cell on the heat strip goes silent, open the Boxes view and pick that box. You see its four EmberSecurity containers, whether each one is running, and when it last reported. Below that are the box's alerts from the last 7 days and its file changes.
That turns a vague worry into a short list of facts. Is the box up? Which container stopped? When did it last speak? The answer to the first question usually tells you who to call.
Counts that move are a signal too
Silence is one thing to watch. A sudden rise is another. When Plasma's edge firewall is switched on, every packet it refuses is counted, and the counts reach EmberCORE through EmberSecurity every 30 seconds. They appear in the Blocked traffic and Threat map views. A count that suddenly rises is usually a device that is new, misconfigured, or not supposed to be there, which is exactly the thing you want to hear about early.
A routine you can run each shift
You do not need a project to get value from this. A shift lead can do it in a couple of minutes.
- Open Command and look at the heat strip before you look at the alert list.
- Count the silent cells. If any are there, open the Boxes view for each one.
- Only then read the queue. An empty queue next to a full set of reporting boxes is good news.
- Work what is there. Acknowledge it, assign it to the person who will follow up, and close it with a note saying what it was and what was done.
Every one of those actions is recorded with who took it and when, and it appears in the audit trail. The next shift reads your note, and an auditor can read it later.
What to do next
Ask the people who run your current monitoring a simple question: what does your screen show when a sensor stops reporting? If the answer is nothing, that is worth fixing no matter what you use.
The EmberSecurity console documentation walks through every view and how to work an alert. If you want to see the heat strip on your own boxes, tell us about your plant and an engineer will walk you through it.