PI to CONNECT change synchronization
- Last UpdatedAug 18, 2026
- 6 minute read
The PI to CONNECT Agent supports synchronizing changes in the Data Archive and PI Asset Framework. Updates are automatically sent to CONNECT without any user interaction.
Data Archive synchronization
The PI to CONNECT Agent signs up for the following updates to stay in sync with the Data Archive:
-
PI point updates
-
Adding a PI point: If PointId was listed in the transfer specification, a stream is created and the data streams to the data store.
-
Updating a PI point name: The stream name is updated.
-
Updating metadata: The corresponding stream metadata is updated.
-
Deleting a PI point: The stream is deleted if the AutoDeleteCloudObject flag is enabled in the transfer settings.
-
-
Digital state updates
-
Update set name/values: The stream values are changed to the updated digital state set/name. See the Data Archive synchronization limitations table below.
-
-
Data updates
-
Point compression updates.
-
Sign-up for PI point updates and digital state updates occurs when the agent starts, while data updates do not begin until the start of a transfer.
Data Archive synchronization limitations
|
Issue |
Restrictions |
|---|---|
|
Toggling compression |
Toggling On to Off: It is possible that the PI point's corresponding stream may be
missing an event near the time when the PI point's compression setting was changed
from On to Off. |
|
Analysis backfilling and rapid succession of data addition and deletion |
There is a possibility for a data gap after an analysis backfill operation due to the nature of rapid data deletion and insertions being performed by the analysis service on the PI point. |
|
Updating digital state |
The stream will store the updated state name and values at time of change. Previously
stored state values are preserved. |
|
Changes to existing events are not captured during agent downtime |
See Agent restart. |
Asset Framework synchronization
The agent performs an indexing of the AF server after agent registration is successful. The indexing caches all known elements and templates along with their attributes. The agent updates this index periodically.
When a transfer starts, the agent builds out the implicit PI points referenced by the element Ids specified in the transfer specification to enable change synchronization.
The supported AF change synchronization events and the result of each change are listed below:
-
Database
-
Database addition: No effect on the existing transfer, but elements from the new database will be available to add if the transfer is edited.
-
Database rename: The name of the database entity is updated.
-
Database deletion: All elements and templates that were also deleted because of the database being removed will propagate and synchronize.1
-
-
Element
-
Updates to Name, Description, Template: The corresponding entity property is updated.
-
Addition or updates to attributes: A stream is created if newly referenced in transfer. The stream relationship of the entity is added or updated.
Note: If an AF attribute references a PI point that does not yet exist (or cannot be resolved) when the transfer is started, creating the PI point or fixing the reference later will not automatically trigger AF change synchronization for that attribute. Create the PI point reference before starting the transfer, or restart/resync the transfer after the reference is created.
-
Deletion of attributes: The stream relationship is removed from the entity. The stream is removed if a point is no longer referenced in the transfer.1
-
Deletion of the element: The entity is deleted.1
-
-
Point attributes
-
Metadata updates (Name, Description, etc.): The entity property is updated to reflect the change.
-
PI point reference updates: The stream relationship of the entity on CONNECT changes to the updated stream. If the previously referenced point is no longer referenced implicitly or explicitly in the transfer, the stream is removed.1
-
New ad hoc stream relationships added directly to the CONNECT entity will not be overwritten by the PI to CONNECT Agent.
-
-
Static attributes
-
Metadata updates (Name, Description, etc.): The entity metadata associated with the attribute is updated.
-
Value updates: The value of the entity property is updated.
-
New ad hoc properties added directly to the CONNECT entity will not be overwritten by the PI to CONNECT Agent.
-
-
Template
-
Updates to Name, Description, Template: The corresponding property is updated on the Entity Component Type.
-
Addition or updates to attribute templates: The stream relationship and properties associated with the Entity Component Type are updated.
-
Deletion of attribute templates: The stream relationship and properties are removed from the Entity Component Type.
-
Deletion of the template: The Entity Component Type is removed.1
-
-
Point attribute templates
-
Metadata updates (Name, Description, etc.): The Entity Component Type property is updated to reflect the change.
-
PI point reference updates: The stream relationship is updated on the Entity Component Type.
-
-
Static attribute templates
-
Metadata updates (Name, Description, etc.): The Entity Component Type property associated with the attribute is updated.
-
Attribute template value type updates: The Entity Component Type property is updated.
-
1 This action occurs only if the AutoDeleteCloudObject flag is enabled in the transfer settings.
PI point type change
When the PI to CONNECT Agent detects that a PI point's type is changed on the source Data Archive after the corresponding stream has been created in the data store, it takes the following actions:
-
The PI Point Type Change Detected message displays next to Current Activity in the Details pane, as shown below:

-
The agent prevents data being sent from the source PI point to the stream until the type is changed to match the corresponding stream type.
-
The agent logs details about the corresponding stream in both the Windows Event Viewer and CONNECT logs.
Causes for PI point type changes
A point change can occur for multiple reasons, including:
-
The source PI point data and type is configured incorrectly. The data and point must be deleted and recreated.
-
The source PI point was misconfigured initially. For example, the point needs to be updated from Float32 to Float64. The data is still relevant and should be kept.
Resume streaming data to a stream after a type change
After you create a stream, its underlying SdsType cannot change. As a result, new data from the PI point in question cannot be stored in the same stream. You can resume streaming data from the PI point to the existing stream by taking the following corrective actions:
-
View the View PI to CONNECT Agent logs in the Windows Event Viewer or Logs to determine which PI point incurred a PI point type change.
Note: The Windows Event Viewer logs are the preferred source of information for PI point type changes.
-
Change the PI point type to match the stream type and then restart the transfer.
To see what types of point coercions are supported in Data Archive, refer to the Allowable point type coercions topic.
If, instead, your goal is to keep the new PI point type and switch to a new corresponding stream type:
-
Delete the affected stream.
-
Restart the transfer.
The agent will recreate the stream using the current PI point type and perform a history recovery/backfill. If there is a large amount of history, the recreated stream can take some time to recover, depending on the volume and throughput.
Agent restart
For performance, the PI to CONNECT Agent tracks the timestamp of the last value sent for each point. When the PI to CONNECT Agent is restarted, it retrieves historical data starting from each PI point's last update time. It then transitions to real-time data collection.
Therefore, if the agent was offline and data was inserted/deleted/substituted for a time range it had already processed, the agent would not capture these values upon restart.