A flat pressure trend may mean a process is stable, a pump is stopped, a sensor has failed, or a display is repeating an old value. Before using that trend to rule out a problem, establish whether the observation is current and whether the source considers it trustworthy.
Stale data is a value that is too old for the decision being made. The acceptable age depends on the signal and task. A daily inventory reading and a rapidly changing safety-related process value should not share an arbitrary freshness threshold.
Follow the value back to its source
Consider a pump temperature that has displayed the same value for two hours. First ask the operator whether the equipment was operating and whether a stable temperature is plausible. Then inspect the value at each available point: the instrument or controller, the source server, the collector and the destination. Follow approved site procedures; do not disturb control equipment to test a dashboard.
For each point, record the value, any available quality or communication status, the time of the observation and the time it was received. A current receipt time can accompany an old measurement if an intermediary repeatedly sends a cached value. Conversely, an unchanged source timestamp may be normal if a source reports only changes. Learn the actual update and timestamp semantics before deciding.
Use several clues, not a flat line alone
| Evidence | Possible interpretation | Next check |
|---|---|---|
| Source value and independent related signals are stable | The process may genuinely be steady. | Compare operating state and expected range. |
| Source status indicates a fault or poor quality | The reading may be unreliable even if a number is present. | Inspect the device and status definition. |
| Source updates but destination does not | A collection, mapping or delivery path may be interrupted. | Check authorized diagnostics and last successful update. |
| Destination receives repeated values with old source time | An intermediary may be forwarding a cached observation. | Identify cache behavior and communication state. |
| Both value and timestamp remain unchanged | Could be a failed path or a change-only update model. | Compare expected heartbeat, status and related signals. |
Quality codes can carry context beyond the numeric value. Their meaning is source-specific; this OPC UA quality overview explains one protocol’s approach. A good status alone does not prove a sensor is physically correct.
Set an actionable freshness rule
For each critical measurement, agree on the expected update behavior, acceptable age, clock basis, responsible owner and response. Some change-driven signals need a separate heartbeat or communication indicator rather than an alarm on an unchanged value. Check that such an indicator actually covers the relevant portion of the path.
Test a safe, approved communication interruption and verify what the application displays. It should distinguish a known stale or uncertain value from a current valid value where the system supports that distinction. Record what happens on recovery and whether missed history is available; never fill a gap with a repeated value without making that treatment explicit.


