Fleet Alerts: What Is P1, What Is P2, and What Can Wait
An alert nobody acts on is not a safety net. It is noise, and noise trains people to ignore the alert that mattered. This sheet sorts the common fleet alerts into three priorities and states what each one is actually asking somebody to do.
Every fleet platform rollout follows the same arc. Week one, everything is switched on, because why would you not want to know. Week three, the supervisor's phone has made four hundred noises, most of them about a vehicle idling at a loading bay it idles at every single day. Week five, notifications are off. Week nine, a vehicle moves at two in the morning and the alert lands in a folder nobody has opened since March.
The fix is not fewer alerts. It is deciding, before you switch anything on, what each alert asks someone to do and how quickly. An alert with no action attached should not exist. An alert whose action is read the report on Monday should never reach a phone.
Three priorities are enough. P1 interrupts a person now, wherever they are, because someone or something is at risk and the window to act is minutes. P2 goes to a named owner during their working hours. P3 belongs in a list that gets worked through rather than pushed. The classifications below are a defensible starting point, not gospel: a cold chain operator and a plant hire company will reasonably disagree about several of them.
Safety and driver alerts
| Alert | Why it matters | Priority |
|---|---|---|
| Impact or crash detected | An accelerometer signature consistent with a collision. Someone may be hurt and unable to call. | P1 |
| Panic button pressed | A driver has asked for help deliberately. There is no version of this that waits. | P1 |
| Stationary in a live lane | Stopped on a motorway or dual carriageway outside any known stopping point. Minutes matter. | P1 |
| Severe over-speed | Well above the limit rather than marginally over it. A live risk, not a coaching point. | P1 |
| Harsh event cluster | Repeated harsh braking in a short window suggests fatigue, distraction, or a vehicle fault. | P2 |
| Driving hours near the limit | The driver is running out of legal time. Actionable while there is still room to replan. | P2 |
| Single over-speed event | One brief excursion over the limit. Real, but the pattern across weeks is what deserves a conversation. | P3 |
| Unidentified driver on a trip | The vehicle moved without a driver tag. Everything measured on that trip belongs to nobody. | P3 |
Security and loss alerts
| Alert | Why it matters | Priority |
|---|---|---|
| Movement outside authorised hours | A vehicle moving at three in the morning from a locked yard. The window to recover it is now. | P1 |
| Towed without ignition | Position changing while the engine is off is not driving. It is a vehicle leaving on somebody else's truck. | P1 |
| Device power disconnected | Permanent power cut at an unusual time or place. Concealment is the most likely explanation. | P1 |
| Sudden fuel drop, engine off | Tank level falling fast while stationary away from a pump. Consumption cannot explain it. | P1 |
| Unexpected geofence exit | An asset leaving a site it should not leave. Urgency depends on how far it can get before anyone notices. | P2 |
| Refill short of the purchase | The card says sixty litres, the tank received forty-five. A discrepancy to investigate, not to chase. | P2 |
| Unscheduled stop on a route | A stop with no drop attached. Sometimes a coffee, sometimes the start of a pattern. | P3 |
| Repeated fuel variance on one vehicle | The vehicle whose books have not balanced for six weeks. Slow, systematic, and worth the arithmetic. | P3 |
Vehicle health alerts
| Alert | Why it matters | Priority |
|---|---|---|
| Red stop lamp active | Engine management has decided the vehicle must stop. Driving on turns a repair into a rebuild. | P1 |
| Low oil pressure | A genuine reading below the safe range. This one ends the shift where the vehicle is standing. | P1 |
| Coolant temperature over range | Overheating in progress. The cost of the next ten minutes is measured in cylinder heads. | P1 |
| Amber warning lamp active | A fault the vehicle can run with. It gets to the depot, then it gets a bay before it turns red. | P2 |
| Charging system fault | An alternator or battery problem. The vehicle finishes the shift and may not start after it. | P2 |
| Particulate filter regeneration required | The filter needs a run it is not getting. Ignored long enough it becomes a forced service and a derate. | P2 |
| Low exhaust fluid level | The vehicle will derate by design when the tank empties. A refill problem with a countdown attached. | P2 |
| Intermittent code recurring | The same fault appearing weekly and clearing itself. It will not fail in the bay, so it needs the history. | P3 |
Compliance and service alerts
| Alert | Why it matters | Priority |
|---|---|---|
| Cold chain temperature excursion | The load is going out of specification. The clock started before the alert did. | P1 |
| Safety defect at walkaround | A driver has recorded something that makes the vehicle unroadworthy. It does not leave the yard. | P1 |
| Roadworthiness test due | A hard legal date. It stops being a diary entry and becomes an off-road vehicle the day it passes. | P2 |
| Service interval reached | Distance or hours have hit the trigger. Booking it this week costs a slot; missing it can cost a warranty. | P2 |
| Licence or qualification expiring | A driver about to become someone you cannot legally roster. Weeks of notice, no ambiguity. | P2 |
| Detention beyond the free period | The vehicle has sat at a customer site long enough to bill for. Worth invoicing, not worth a phone call. | P3 |
| Defect open past its close-out date | Something was reported, recorded, and never fixed. This is the alert that turns up in an investigation. | P3 |
Device and data health alerts
| Alert | Why it matters | Priority |
|---|---|---|
| Tracker silent during a shift | A vehicle known to be working has stopped reporting. Absence of data is itself data. | P2 |
| Fuel sensor flatlined | A tank level that has not moved through a full shift. Every fuel figure on that vehicle is now fiction. | P2 |
| No position fix while moving | The unit is online but has no position. Aerial, obstruction, or something in the cab drowning the signal. | P2 |
| Ignition signal implausible | Trips and idle time are derived from this. If it is wrong, so is everything computed on top of it. | P2 |
| Device offline over a weekend | A parked vehicle gone quiet. Usually a flat battery, occasionally the first sign of something else. | P3 |
| Firmware out of date | Units drifting behind the rest of the fleet. Not urgent, and it accumulates until a rollout needs a workshop visit. | P3 |
Frequently asked questions
Few enough that each one gets read. There is no correct number, but there is a reliable test: if your supervisor cannot recall what the last alert was about, the volume is already too high and the system has stopped working. Start with the P1 list only, run it for a month, then add P2 categories one at a time with a named owner for each. A fleet that switches everything on at once ends up with everything switched off.
Yes, and that is exactly why the P1 list has to be short and defensible. If it wakes someone at two in the morning for a vehicle idling at a depot, the escalation path will be quietly disabled within a fortnight and nothing will wake anyone for a real event. Before adding an alert to P1, answer one question: who is going to do what, in the next ten minutes, at three in the morning. No answer means it is not P1.
By being straight about what is measured and by using it consistently. The complaint is rarely about the data; it is about the data appearing only when somebody is in trouble. If harsh braking events are used to defend a driver in a disputed accident, and to fix a junction where everyone brakes hard, the objection tends to fade. If they only ever appear in a disciplinary, it will not, and the drivers will be right. Publish the thresholds, do not change them quietly, and tell people what is on before you switch it on.
You can, and you should, because an idling threshold that suits a city multi-drop is wrong for a plant fleet running hydraulics. Two failure modes to watch. The first is tuning a threshold until the alerts stop, which is not the same as fixing anything. The second is changing thresholds mid-quarter, which makes your own trend data uncomparable: the drop in harsh braking events will look like a safety win and will be an administrative one. Set them, record why, and review on a schedule rather than after a bad week.
There are things we will not turn into an alert, because the underlying signal cannot carry the claim. Fatigue inferred from steering inputs alone is one: a tired driver and a driver on a poor road surface produce similar traces, and an alert accusing somebody of being unsafe needs better evidence than that. The general rule is that an alert should describe what was measured, not diagnose a cause. Fuel level fell forty litres while stationary is a fact. This driver stole fuel is a conclusion, and it is yours to reach.
See these metrics on your own fleet
GPS Fleet Tracking Software reports every one of these numbers from live operational data.