| Measurement |
Description |
Measurement
Unit |
Interpretation |
| Pct_conns_used |
Percentage of
database connection allocated to a connection pool that are in use |
Percent
|
When this value
reaches 100%, connections to the database will either have to wait for
access or will time out. Consider increasing the size of the
connection pool in this case. A very high percentage of use can also
result if one or more of the applications that use the connection pool
are not releasing the connections after their use. |
| Pending_conns |
Number of
requests for connections to the database that are currently pending |
Number
|
A high value of
pending connections is indicative of a bottleneck during database
access. Reasons for this could either be that the connection pool
capacity is insufficient, or that the database server has slowed down,
causing requests to take longer to execute their queries. |
| Max_conns |
The maximum number of connections configured for a JDBC connection
pool
|
Number
|
|
| Current_conns |
The current
number of connections in use for a JDBC connection pool |
Number
|
|
| Pool_availability |
The current state
of a database connection pool – whether it is available or not |
Percent
|
A value of 100
denotes that the pool is available; a value of 0 denotes
unavailability. |
| Failed_reconnects |
The number of
cases during the last measurement period when a connection pool
attempted to refresh a connection to the database and failed |
Number
|
Failures could
happen if the database is unavailable, or, the connection was
terminated. |
| Connections_to_pool |
The number of
JDBC connections added to the pool in the last measurement period |
Number
|
|
| Avg_conn_delay |
Avg. time taken
to get a connection from the database (in seconds) |
Secs
|
An increase in
this metric could indicate a bottleneck in the database tier. |
| Leaked_conns |
Number of
connections tagged as a leaked connection during the last measurement
period. A leaked connection is a connection that was checked out from
the connection pool but was not returned to the pool by calling
close(). |
Number
|
A non-zero value
indicates that the application may not be releasing connections after
use. This can result in inefficient management of the database
connections in the pool. |
| Active_conns_high_mark |
The high water
mark of active connections in a pool |
Number
|
Note the changes
in the high water mark. This can provide an idea of the times when the
JDBC pool was most heavily used. |
| Waiting_conns_high_mark |
The high water
mark of number of waiters for a connection from the pool |
Number
|
Note the changes
in the high water mark. This indicates periods when the database
connection pool could have been a bottleneck. |
| Note:
The
Failed_reconnects, Avg_conn_delay and Leaked_conns measures will
always report 0 for WebLogic 6.0 with Service Pack 2.0. |