Limitations of PI to CONNECT

Restrictions for PI point transfers

• Point must belong to the classic or base point class.

• Point type must be Float16, Float32, Float64, Int16, Int32, Digital, Timestamp, or String.

• Must not be a future PI point.

High availability (HA) & collectives

• PI to CONNECT does not support the high availability (HA) feature of failover between collective nodes. Connecting to a Data Archive collective is supported, but the agent must explicitly point to either the primary or secondary server. The agent will not fail over between members of the collective if a connection to the configured endpoint is lost.

• AF elements or attributes that reference a collective name instead of one of the nodes are not supported.

• AF Collectives are supported natively by PI to CONNECT.

PI to CONNECT Agent registration

• Multiple PI to CONNECT Agents can connect to and transfer data from the same Data Archive if the agents are installed on different machines and assigned to different PI to CONNECT service instances. Only one agent can be installed on each computer.

Data Archive

• PI to CONNECT does not write events back to Data Archive from CONNECT.

Stream indexes

• Does not support multiple values at a given index.

• If a Data Archive tag has multiple values at a given timestamp, CONNECT will store the first value returned.

Custom UOM data

• Custom units of measure (UOMs) that are not one of the predefined UOM classes in CONNECT are not supported. During the transfer of AF element data, AF elements with custom UOMs will not have their corresponding entity's UOM property set.

PI Analysis Service

• If you use PI to CONNECT with the PI Analysis Service, be sure to save analysis outputs to PI points.

Multiple UOMs

• If multiple attributes reference the same PI point but have different UOMs, a stream UOM is not transferred.

AF data limitations

• For limitations on the transfer of AF data, see AF data that can be transferred.