Documentation forDatabase Performance Analyzer

Db2 metrics collected by DPA

The following sections list the metrics that DPA collects for Db2 self-managed database instances.

  • Learn how to view these metrics and change the thresholds.
  • For detailed information about resolving issues, click the next to the metric on the Resources tab. The Information link is not available for all metrics.
  • If the database instance runs on a virtual machine (VM), metrics collected for the VM are described in VM metrics collected by DPA

Connections

Metric Description
Connected Devices The number of distinct client machines connected (even if the connection is idle).
Connected Users The number of distinct users connected (even if the connection is idle).
Sessions The number of sessions connected (even if the connection is idle).

CPU

Metric Description
Core Count The number of cores used by the instance.
O/S CPU Utilization The percentage of CPU being used for the entire system (includes all databases on this machine). Potential solutions to a CPU bottleneck are to reduce the server load (tune those queries), get faster CPUs, or get more CPUs.

Disk

Metric Description
DB Physical I/O Rate

The number of read and write operations performed to/from disk every second. If this is high, drill in to the DPA Trends charts and review queries waiting on physical read and write wait events.

Db2 might supply this value only if the DFT_MON_BUFPOOL monitoring switch is ON.

DB Physical Read Rate

The number of reads performed from disk every second. If this is high, drill in to the DPA Trends charts and review queries waiting on physical read wait events.

Db2 might supply this value only if the DFT_MON_BUFPOOL monitoring switch is ON.

DB Physical Write Rate

The number of writes performed to disk every second. If this is high, drill in to the DPA Trends charts and review queries waiting on write wait events.

Db2 might supply this value only if the DFT_MON_BUFPOOL monitoring switch is ON.

Memory

Metric Description
DB Buffer Pool Hit Ratio

The rate at which Db2 finds the data it needs in memory rather than having to read from disk. By itself, the buffer pool hit ratio is not very meaningful except for databases with undersized memory settings. Db2 might supply this value only if the DFT_MON_BUFFERPOOL monitoring switch is ON.

Tuning queries and performing index optimization is the best way to increase buffer pool hit ratios.

DB Catalog Cache Hit Ratio The percentage of time when Db2 looks for an execution plan in the catalog cache and finds it. A low hit ratio indicates the catalogcache_sz parameter should be increased.
DB Package Cache Hit Ratio The percentage of time when Db2 looks for an execution plan in the package cache and finds it. A low hit ratio indicates the pckcachesz parameter should be increased.
O/S Memory Utilization The percentage of system memory being used for the entire system (includes all databases on this machine). If this is high, you might need to increase the amount of physical RAM in the server, reduce the load on the server, or change your server memory configuration.
Virtual Memory Utilization The percentage of virtual memory being used.

Network

Metric Description
DB Round-trip Time The round-trip time when running "select 1 from sysibm.sysdummy1" against the database specified during registration (includes network time but not connect time). If this is high, contact your network administrator to understand network latency.

Sessions

Metric Description
DB Blocked Sessions The number of sessions that are waiting on lock waits because another session is using a needed resource.
DB Connections Currently Executing The number of sessions that are actively performing work or waiting for a resource (excludes idle sessions).
DB Transaction Rate

The number of transactions being executed every second:

commit_sql_stmts + int_commits + rollback_sql_stmts + int_rollbacks

Waits

Metric Description
Total Instance Wait Time The total wait time for the instance.