Performance monitoring for PI Buffer Subsystem

You can use Windows Performance Monitor and PI performance counters to monitor buffering health and status across the organization. In contrast, you can use Buffering Manager to monitor buffering on one client or interface computer.

The three performance objects provided with PI Buffer Subsystem 4.3 and later include counters that allow you to monitor buffering health and status at three different levels. Each object is described below, along with its key counters.

PI Buffer Subsystem object (global status)

The PI Buffer Subsystem object has counters for monitoring global status, which includes all buffering from a given client or interface computer. For example, the global counter Total Events Sent/sec tells you the number of events sent to all servers receiving buffered data from this computer.

Key global performance counters include:

• Health: the overall health of PI Buffer Subsystem, including the health of all servers receiving buffered data (0 = OK, 1 = Warning, 2 = Error, 3=Critical).

• Events Received/sec: the number of events per second received by PI Buffer Subsystem.

• Queue> Capacity: estimated time (in seconds) until data loss if all servers were disconnected.

• Total Events Sent/sec: the number of events per second sent to all servers receiving buffered data.

For more details and a complete list of global performance counters for PI Buffer Subsystem, see PI Buffer Subsystem statistics.

PI Buffered Logical Servers object (logical server status)

The PI Buffered Logical Severs object has instances and counters you can use to monitor the status of buffered logical servers. A logical server is either a standalone Data Archive server or a Data Archive server collective, which appears as one server to end users. When users connect to a collective, they may not be aware that they are actually connecting to one of its members.

For example, assume you are sending buffered data to the collective COL1 and the server SRV3 shown below. To monitor the status of COL1, you can add counters from the PI Buffered Logical Servers object for the instance COL1. To monitor the status of SRV3, you can add counters for the instance SRV3.

Key logical performance counters include:

• Health: the overall health of this logical server, including the health of all physical servers that comprise this logical server (0 = OK, 1 = Warning, 2 = Error, 3=Critical).

• Events Received/sec: the number of events per second received by PI Buffer Subsystem for this logical server.

• Events Rejected: the number of events rejected by PI Buffer Subsystem for this logical server.

• Points with Post Error: the number of points for which PI Buffer Subsystem failed to post the last event (sum of Points with Post Error of all physical servers that comprise this logical server).

• Total Events Sent/sec: the number of events per second PI Buffer Subsystem sent to all physical servers that comprise this logical server.

• Total Queued Events: the total number of events in all buffer queues for physical servers that comprise this logical server.

For a complete list of logical performance counters for PI Buffer Subsystem, see PI Buffered Logical Servers statistics.

PI Buffered Physical Servers object (physical server status)

The PI Buffered Physical Severs object has instances and counters you can use to monitor the status of individual buffered servers, including both collective members and standalone Data Archive servers.

While certain counters are available for both PI Buffered Logical Servers and PI Buffered Physical Servers, each object also has its own unique counters. For example, the physical counters State> Is Connected and State> Is Sending OK make sense only for individual physical servers, while the counters Total Events Sent/sec and Total Queued Events may be useful at both the logical and physical levels.

Key physical performance counters include:

• Health: the health of buffering for this physical server (0 = OK, 1 = Warning, 2 = Error, 3=Critical).

• Points with Post Error: the number of points for which PI Buffer Subsystem failed to post the last event to this server.

• State> Is Sending OK: whether PI Buffer Subsystem is sending or ready to send data to this server (0 = No, 1 = Yes).

• Total Events Sent/sec: the number of events per second PI Buffer Subsystem sent to this physical server.

• Total Queued Events: the total number of events in all buffer queues for this physical server.

For a complete list of logical performance counters for PI Buffer Subsystem, see PI Buffered Physical Servers statistics.