Using the dashboard health summary
You can now use the dashboard view that you created to monitor the health of replication between the LONDON source database and the DALLAS and TOKYO target databases. This section gives you a tour of your london.asn dashboard view.
The Summary page has three main parts:
- A topology graphic that depicts each configuration.
- A latency and throughput summary for one of the Q Capture-Q Apply pairs.
- A Replication Configurations table in which each row shows rolled-up states for queues, subscriptions, latency, exceptions, and messages. The following notation in the Q Capture (Send queues) and Q Apply (Receive queues) columns stands for "program state (queues state, rolled up across all queues)":
- In the Replication Configurations table, click on the unidirectional icons
to the left of the two LONDON.ASN Q Capture entries to expand your view of the
Figure 12. Expanded configurations table
The expanded table includes a row for each send queue-receive queue pair. In this case, the send and receive queue are the same local queue. The rows include the following information:
- The status of the send queue and receive queue
- The number of Q subscriptions for each queue, and an icon that shows their status
- The average end-to-end latency for transactions that were sent on the queues during the last refresh interval
- The number of data exceptions that were written to the IBMQREP_EXCEPTIONS control table for Q subscriptions that use the queues
- Icons that notify you of any error messages that the Q Capture or Q Apply programs generate
Note: Some of the values in your Replication Configurations table are likely to be different from the screen capture above, especially latency and the number of exceptions. The latency depends on a number of factors, including the speed of your system, and the number of exceptions will differ depending on the length of time between starting replication and creating the dashboard view.
- In the upper-right area of the Summary page, notice the graphical latency
and throughput summary for the LONDON.ASN > DALLAS.ASN configuration:
Figure 13. Latency and throughput summary
The moving chart tracks two important measures of replication performance for one Q Capture-Q Apply pair:
Table 4. Replication performance measures Measure Description Log latency The average time that elapses between transactions being committed at the source table and the Q Capture program reading the transaction commit statement in the recovery log. This measurement, in seconds, gives you a rough idea of how "caught up" Q Capture is in reading the log. Q Apply throughput The average number of rows per second that the Q Apply program applied to the target table.
To view charts for both configurations at the same time, replace the Topology graphic -- click on the Change throughput summary icon above the topology: The topology view is a useful way of quickly seeing which servers are connected and how. For active monitoring, you might prefer to see both charts.
- In the Replication Configurations table, look for the row that identifies
SEND_RECVQ2_LON_DALLAS. Note that in the Q subscriptions column, both Q
subscriptions are active. To get more details about the two Q subscriptions,
click the column that contains the number 2 next to a green diamond:
Figure 14. Drilling down on Q subscription state
The Q subscription details window that opens provides more detail on the status of all Q subscriptions that use this send queue-receive queue pair.
- In the Q subscription details window, click the number 2 next to
"Active:", as Figure 15 illustrates:
Figure 15. Q subscription details window
The Subscriptions page opens with a table that provides more details about the Q subscriptions and their state. From this page you can start, stop, and reinitialize Q subscriptions and publications, and also indicate that a manual load of the target table is complete. You can also use the drop-down boxes to view Q subscription status for another configuration (Q Capture > Q Apply), or for a different send queue-receive queue within the same configuration:
Figure 16. Q subscription drill down
You can use the health summary table to drill down on programs and queues in addition to Q subscriptions, and the dashboard provides the same drill-down capability for latency statistics, exceptions, and messages. Click any table cell or link in the summary table to drill down into more detailed information.
- Note that the dashboard view is divided into five tabs:
Figure 17. Tabs on dashboard view
In addition to the Summary tab, the following tabs give you the ability to monitor and manage other elements of your replication environment:
Table 5. Dashboard view tabs Tab Description Live graphs Provides graphical, near-real-time latency and throughput information for each send queue-receive queue pair. Programs Shows the status of the Q Capture and Q Apply programs and current values of their operation properties. You can also change property values. Queues Shows the status of queue managers and queues, including queue depth. You can also view and change queue map properties. Subscriptions Shows the status of Q subscriptions or publications.
- Select the Live Graphs tab. Each block of four moving graphs shows
statistics from one send queue-receive queue pair (the latency and throughput
graph on the summary page is for a Q Capture-Q Apply pair). The Live Graphs
tab provides statistics on a more detailed level than the graph on the Summary
page, and enables you to see latency and throughput data for each specific
send queue-receive queue pair rather than just the rolled-up information that
you saw in the Summary page.
The graphs provide these additional statistics:
Table 6. Additional graph statistics Measure Description End-to-end latency The average elapsed time in seconds between transactions being committed at the source database and being applied at the target. This is the same measure for which you can set a threshold when you create or change a dashboard view. Q Capture throughput The number of rows per second that the Q Capture program puts on the send queue.
Figure 18. Live Graph page
You can enlarge the graphs to 150 percent or 200 percent, or change which live graphs are displayed by clicking on the Change live graphs control in the upper-right area:
You can also maximize the number of live graphs that display by clicking on the Open new window control to open the live graphs in a separate window:
Or click on the Legend icon to open a legend window that describes what the colors and lines in the graphs mean:
Figure 19. Legend for live graphs
You have now had a tour of the overall structure of a dashboard view. The next section will show you how to use the dashboard to track and analyze replication latency between the source and target tables.