PI Nexus+ / Reference
PI Interface Issues and Runtime
This reference lists the PI Interface issues, the runtime statuses and the runtime thresholds. The User Guide chapter PI Interfaces links here.
Overview
This reference lists the PI Interface issues, the runtime statuses and the runtime thresholds. The User Guide chapter PI Interfaces links here.
Configuration issues
These decide the Configuration column. They appear on the Issues page under PI Interfaces. All except the last are raised on the PI Point.
| Issue | Severity | Raised when | Typical fix |
|---|---|---|---|
| Unmapped PI Point Interface | Warning | The point has a PointSource and an interface ID (Location1), but no interface in the Module Database matches both | Correct the point's PointSource or Location1, or register the interface |
| Ambiguous PI Point Interface | Error | More than one interface matches the point's PointSource and interface ID | Give each interface a unique ID per PointSource |
| Invalid OPC Processing Mode | Warning | Location3 of an OPC point is not a supported processing mode | Correct Location3 |
| Undefined Interface Scan Class | Error | Location4 names a scan class the mapped interface does not define | Add the scan class or correct Location4 |
| Mixed Advised and Polled Scan Class | Warning | One scan class of an OPC interface contains both advised and polled points (raised on the interface) | Move advised and polled points to separate scan classes |
Configuration and Data Health are refreshed by the PI Interface Health step of every PI Point inventory and health refresh scan, by a PI Interfaces scan, and by Check health on one interface. Interface definitions, scan classes and mappings come from an inventory scan.
Runtime issue
| Issue | Severity | Raised when |
|---|---|---|
| PI Interface Runtime Health | Error when Runtime is Error; Warning when Runtime is Warning or Unknown | The runtime evaluation is Warning, Error or Unknown for long enough (see thresholds) |
The message names the cause:
| Message | Cause |
|---|---|
| The interface heartbeat has not updated within its expected interval. | Heartbeat older than 2 minutes (Warning) or 5 minutes (Error) |
| The interface heartbeat reports bad quality or a shutdown state. | Heartbeat point has bad quality or a shutdown digital state (Error) |
| The interface device status reports an error state. | UniInt device status in error, such as Device(s) in error or Intf Shutdown |
| The interface device status is not Good. | Any other non-Good device status, or an empty one |
| Runtime telemetry could not be read from PI. | Reading the runtime points failed |
| A scan class scan time reaches its configured scan period. | Scan time at 80 % (Warning) or 100 % (Error) of the scan period |
| A runtime performance counter is outside its expected range. | Missed or skipped scans above 0 % (Warning) or 5 % (Error) |
| Runtime telemetry has not produced a usable value. | The heartbeat has no timestamp, or no runtime value is usable |
Throughput, I/O rate and process counters are shown as evidence only and never raise an issue. A suppressed runtime issue sends no email.
Runtime statuses
| Status | Meaning |
|---|---|
| Good | Heartbeat current and device status Good |
| Warning | Heartbeat older than 2 minutes, a warning device status, or a performance finding (see the causes above) |
| Error | Heartbeat older than 5 minutes, bad quality, an error device status or a shutdown state |
| Unknown | The mapped core signals could not be read, or have not produced a usable value yet |
| Setup required | Runtime Monitoring is on, but no heartbeat or device-status PI Point is mapped; set the UniInt marker in the health point's Extended Descriptor or map it in Monitoring setup |
| Not Monitored | Runtime Monitoring or Interface Tracking is off, or the archive is disabled |
A Warning or Error from the heartbeat or the device status must be seen in 2 polls in a row before the status changes. Recovery to Good shows on the first healthy poll; a read failure shows at once. Heartbeat age is measured against the PI Data Archive's clock. When the runtime poll stalls, statuses keep their last value and the page says monitoring is not refreshing.
Device status values
The code before the first | of the UniInt [UI_DEVSTAT] value decides the status.
| Reported value | Status | ||||
|---|---|---|---|---|---|
Good, 0, `0 \ | Good` | Good | |||
| `1 \ | Starting, 2 \ | Connected/No Data (legacy); 10 \ | Connected / No Data, 30 to 70, 90 \ | Starting` (UniInt 4.7) | Warning |
| `3 \ | 1 device(s) in error, 4 \ | Intf Shutdown (legacy); 95 \ | Device(s) in error, 99 \ | Intf Shutdown` (UniInt 4.7) | Error |
System states Intf Shut, I/O Timeout, Comm Fail, Bad Input, Configure, Shutdown | Error | ||||
System states No Data, Scan Off, Pt Created, and any unrecognized value | Warning |
An interface writes Intf Shut on a clean shutdown only when it runs with /stopstat. A crash writes nothing, so the heartbeat age is the main shutdown signal.
Runtime thresholds
These are fixed in the product.
| Setting | Value |
|---|---|
| Runtime poll interval | Interface and adapter runtime poll in the scan automation settings (see Scan Automation in the Administration Guide): every 1, 2, 5, 10 or 15 minutes, default every minute |
| Heartbeat Warning / Error | Older than 2 / 5 minutes |
| Polls before a core finding counts | 2 in a row |
| Polls before a performance finding counts | 3 in a row |
| Scan time Warning / Error | 80 % / 100 % of the scan class period |
| Missed or skipped scans Warning / Error | Above 0 % / above 5 % |
| not refreshing banner | No successful poll for 3 poll intervals, at least 10 minutes |
Runtime emails
Configured under Admin > Notifications (see Notifications in the Administration Guide). An open alert is not repeated at the same severity. A rise to a higher severity (Unknown, then Warning, then Error) sends one more message marked escalated. A drop to a lower severity waits for the recovery message.
