
Chapter 11: Runtime Monitoring
Using the Runtime Monitor
The Runtime Monitor allows you to monitor server events and
statistics, which may help you anticipate and prevent server problems.
Starting the Runtime Monitor
Double-click the Servers icon.
Double-click the server you want to monitor.
The Runtime Monitor can connect to other EAServer instances
via an IIOP listener. The server configuration identifies the host
and port number to which the Runtime Monitor attempts to connect.
Click the Runtime Monitor icon to display these folders:
Packages Monitor events and statistics for a specific package, or for
all packages on the server. See “Monitoring component activity” for more information.
Connection caches Monitor a specific connection cache or statistics for all
caches, as described in “Monitoring connection caches and managed connection factories”.
Managed Connection Factories Monitor connection factories or statistics for all connection
factories, as described in “Monitoring connection caches and managed connection factories”.
Network Monitor protocol-specific session information, as described
in “Monitoring network connections”.
Instance Pools Monitor component instance pool activity, including the total
number of instances in each pool, and the number of instances for
each component assigned to the pool.
Active Component Models Highlight this folder to display the component models that are
in use. Information about each model is displayed in the right pane.
If you are running Java or EJB components, you can determine
the version of the server’s Java virtual machine.
If you are running PowerBuilder components using a version
9.0.1 (build 6514) or later PowerBuilder Virtual Machine (PBVM),
you can determine the version and build number of the PBVM that
is running the component.
Statistic counters Some monitoring folders display counters,
which display a monitored statistic, such as component invocations
or HTTP requests. When you highlight the folder, the current values
of the counters display on the right side of the window. There are
several types of counters:
Snapshot
counters represent values at a particular point in time
that are likely to either increase or decrease. For example, number
of sessions, number of instances active, number of active connections,
and so on. Snapshot counters do not display information when set
at the per-second rate; instead they display “N/A.”
Cumulative counters represent
values that always increment and never decrement. For example, number
of invocations, number of connections opened, number of network
requests, bytes read and written, and so on. Cumulative counters
display both per-second and counter information.
Peak maximum counters represent
the total value since starting the server.
Last maximum counters represent
the total value since starting the runtime monitor, and is not useful
when there are multiple monitoring clients.
Refreshing displayed statistics You can refresh or change the counters view by choosing any
of these options from the File menu:
Refresh Refresh the display to obtain the latest counters.
Per-Second Display the per-second values.
Counter Display the accumulated value.
View Values View the values in a separate dialog that automatically refreshes
according to the refresh rate, which you can select from a drop-down
list. If the group contains subgroups, expand those subgroups until you
see the values.
Monitoring component activity
The Packages folder shows statistics about component invocations.
You can view statistics for individual packages or for all packages
installed in the server, including those installed indirectly as
part of an installed application. Table 11-1 describes the counters.
Table 11-1: Component monitoring counters
Counter name
|
Description
|
Method invocations
|
Cumulative count of method invocations
since the server was started.
|
Instances active
|
Snapshot count of active component instances,
that is, instances currently bound to a client session.
|
Instances pooled
|
Snapshot count of pooled component instances.
|
Transactions completed
|
Cumulative count of transactions completed.
|
Transactions rolled back
|
Cumulative count of transactions rolled
back.
|
Last maximum method invocations
|
Maximum count of method invocations since
starting the counter view.
|
Last maximum instances active
|
Maximum count of active component instances
since starting the counter view.
|
Last maximum instances pooled
|
Maximum count of pooled component instances
since starting the counter view.
|
Last maximum transactions completed
|
Maximum count of transactions completed
since starting the counter view.
|
Last maximum transactions rolled back
|
Maximum count of transactions rolled
back since starting the counter view.
|
Peak maximum method invocations
|
Maximum method invocations since the
server was started.
|
Peak maximum instances active
|
Maximum active instances since the server
was started.
|
Peak maximum instances pooled
|
Maximum pooled instances since the server
was started.
|
Peak maximum transactions completed
|
Maximum transactions completed since
the server was started.
|
Peak maximum transactions rolled back
|
Maximum transactions rolled back since
the server was started.
|
Monitoring connection caches
and managed connection factories
The Connection Caches and Managed Connection Factory folders
show statistics for cached database connections. You can display
statistics for all caches or connection factories, or for each one
individually. Table 11-2 describes
the counters that display.
Table 11-2: Connection cache and managed
connection factory statistics
Counter name
|
Description
|
Connections open
|
Snapshot count of open (but not necessarily
active) connections.
|
Connections active
|
Snapshot count of active connections.
|
Connections opened
|
Cumulative count of the number of connections
opened since the server started.
|
Connections closed
|
Cumulative count of connections closed
since the server started.
|
Waited connection requests
|
Cumulative count of connection requests
that required a wait because the maximum number of connections were
in use. If this number is large, consider increasing the connection
pool size.
|
No wait connection requests
|
Cumulative count of connection requests
that required no wait.
|
Forced connection requests
|
Cumulative count of forced connection
requests, that is, connections that were in excess of the configured
pool size, but allocated immediately. If this number is large, consider
increasing the connection pool size.
|
Total connection requests
|
Cumulative count of the total number
of connection requests since the server started.
|
Last maximum connections active
|
Maximum count of active connections since
starting the counter view.
|
Peak maximum connections active
|
Maximum count of active connections since
the server started.
|
Monitoring network connections
The network connections folder allows you to monitor HTTP
and IIOP connections.
HTTP and IIOP statistics
To view HTTP statistics, expand the HTTP folder, then highlight
the HTTP Statistics folder to display the counters described in Table 11-3.
To view IIOP statistics, expand the IIOP folder, then highlight
the IIOP Statistics folder to display the counters described in Table 11-3.
Table 11-3: HTTP and IIOP statistics counters
Counter name
|
Description
|
Sessions
|
Snapshot count of active sessions. This
also represents the number of threads serving clients for this protocol.
|
Requests
|
Cumulative count of requests processed
since the server started.
|
Bytes read
|
Cumulative count of bytes read since
the server started.
|
Bytes written
|
Cumulative count of bytes written since
the server started.
|
Last maximum sessions
|
Maximum number of sessions since starting
the counter view.
|
Last maximum requests
|
Maximum number of requests processed
since starting the counter view.
|
Last maximum bytes read
|
Maximum number of bytes read since starting
the counter view.
|
Last maximum bytes written
|
Maximum number of bytes written since
starting the counter view.
|
Peak maximum sessions
|
Maximum number of sessions since the
server started.
|
Peak maximum requests
|
Maximum number of requests since the
server started.
|
Peak maximum bytes read
|
Maximum number of bytes read since the
server started.
|
Peak maximum bytes written
|
Maximum number of bytes written since
the server started.
|
Maximum threads
|
The thread limit for clients of this
protocol, which equals the maximum number of client sessions supported
by the server configuration.
|
Connected users
The Connected Users folder under the IIOP folder displays
the thread ID, host name, and host IP address for each IIOP client
connection.
HTTP history
The HTTP History/Hits folder allows you to monitor
recent request frequencies for several time intervals, as listed
in Table 11-4. To
enable monitoring of these statistics, expand the Runtime Monitoring/Network/HTTP History/Hits
folder, then choose File | Start Data Collection. You can
refresh the history view by choosing View | Refresh Folder
or by pressing the F5 key.
Table 11-4: HTTP history counters
Counter name
|
Description
|
Avg hits/second last period
|
The hits-per-second rate, averaged over
the specified time period. The value “Not enough data” means
the specified time period has not elapsed since you enabled data
collection, or that collection is not enabled.
|
History collection enabled?
|
A value of True indicates that history
collection is enabled.
|
Copyright © 2005. Sybase Inc. All rights reserved.
|
|