Welcome to your guide!

Real-Time Data Replication Test Drive

Monitoring

The HVR hub gathers all information about the jobs it manages. In this section you will explore various ways to monitor data replication activity.

Topology

The term Topology is used to indicate an overview of all data replication flows managed by the hub. In a single overview it shows the direction of the data flows, an indication of data volumes, a relative indication of latency, as well as whether jobs are running. The chart is interactive to retrieve more information such as actual current latency and volume information.

1. To get to the Topology, navigate to the node Scheduler in the tree.

2. Pull up the context menu and select Topology.

3. The Topology is displayed in a browser window, which at present can only be shown on the machine running the HVR GUI.

4. Inspect this chart carefully, and close the browser when you are done.

More on this topic

1. Use the LEGEND drop-down on the top right to get a feel for the information displayed on the screen.

2. The settings cog to the right of LEGEND can be used to change some of the information displayed in the chart. Note that, at present, these settings reset when you close the browser and return to the Topology.

3. Click on the lines, and then the arrows to get numerical data from the data flow.

4. In the box on the bottom left that shows the high-level status for your hub, click the mins button to get aggregate statistics updated every minute.

Statistics

As data flows between source(s) and target(s), HVR keeps detailed logs. Information is then harvested from these logs to publish historical metrics as time-series statistics about the replication flows. The time series use fixed intervals, either minute, 10 minute, hourly or daily intervals, with metrics aggregated to these time intervals (such as total number of changes, or minimal/maximum latency for the interval, etc.).

1. To access Statistics, in the HVR GUI, navigate to the Scheduler node right click to bring up the context menu and select Statistics.

2. Like the Topology view, Statistics are shown in a browser. In fact, using the bar on the left you can switch between Topology and Statistics (and Events, which you will explore later).

Note that the Statistics have a granularity defaulting to hourly with a default window of seven days.

3. Since replication just started on this hub, and to show more detail, use the drop-down on the top right to switch the granularity to minute.

4. To zoom into the chart, either use the time selector, or hover the pointer over one of the charts until a + (plus) sign appears.

5. Horizontally select the portion of the chart you want to zoom into.

By default, HVR shows the most commonly accessed metrics. (i.e. Latency, Captured Changes Counts, and Integrated Changes broken down by Table/Channel.)

6. Take some time to inspect the statistics, and explore what is available:

• More metrics/statistics are available under the + (plus) button below the Time Selector on the right.
• Reported statistics can be broken down differently.

7. Select or de-select statistics to focus on others, or to have HVR automatically adjust the scale so you have access to more details

8. Hover the pointer over the charts to see a pop-up with the detailed values across all visible charts.

9. Close the browser window when you are done exploring the statistics.

Log Files

1. To inspect a log, use the tree in the HVR GUI to navigate to the Scheduler nodes, or jobs, below the Scheduler.

2. Use the context menu to select View Log. By default, the log will appear on a new tab at the bottom of the HVR GUI, keeping track of the latest changes flowing into the log.

3. Right click on orcl2mdbpg-cap-orcl to bring up the context menu and select View Log. This will allow you to see the capture log for changes captured from the Oracle Database.

4. Notice that changes continue to get added to the log.

5. You can pause the ongoing flow of incoming messages by using the pause button at the top left of the lower pane. Also, note the full path to the log file should you have a need to inspect the log file directly.

6. Explore the other functions that the lower pane offers to inspect the log, and close the pane when you are done.

More on this topic

In some cases, such as an error, there is a need to inspect a log file. HVR stores logs at different levels: per Job, per Channel, and for the entire Scheduler. At every higher level, the logs are simply aggregated from the lower levels (e.g. the Channel log simply consists of the logs for all Jobs in the Channel).

Alerts and other Maintenance Tasks

HVR provides alerts through Maintenance Tasks that are scheduled outside of HVR to regularly check the health of the system, current latency, and whether there have been errors or warnings since the most recent check. Maintenance Tasks are also used to keep the system healthy by rolling over the log file (and keeping a backup for some number of days) and purging old journals that may have been kept during data replication.

1. Select the Scheduler node in the context menu in the tree of your hub.

2. Select Maintenance Tasks on the pop-up in order to access the dialog and configure alerts and maintenance tasks.

3. Inspect the options for alerts and maintenance tasks on the Maintenance Tasks dialog and close the window when you are done.

More on this topic

HVR is designed to automatically recover from interruptions to data replication and will always resume where it left off. As a result, once configured, HVR can be left alone to manage the replication and periodically you may access HVR to look at the Topology chart or inspect Statistics.

However, there can be scenarios in which replication stops and cannot be resumed without intervention. Examples that could cause replication to fail are: the network became unavailable for an extended period of time, a system crashed due to a faulty part, or a bug in software (either HVR’s or the technologies with which HVR integrates).

Many organizations have Service Level Agreements (SLAs) on data replication and administrators would want to be notified when there are major issues that may result in SLAs getting failed.

  • You can create as many tasks as needed, and tasks can be filtered by channel or by location. For example, when creating an alert, some channels may have lower latency SLAs than others so you create separate tasks to verify these.
  • Alerts can be sent out through email, on Slack, or integrated via SNMP into another monitoring solution.
  • With these maintenance tasks checking in on the health of the HVR setup, including the scheduler, they are scheduled using crontab on Linux (or Unix) which is the default scheduler on this operating system. On Windows the tasks are scheduled in the Windows scheduler.

Checkpoint

  • With Replication still running you should now have a feel for how HVR can be used to monitor ongoing replication and keep track of SLAs.
  • The Topology is a great visual to get a quick overview of all data replication managed by the hub, with Statistics showing aggregate information across the hub with many options to drill into the details.
  • Alerts can be configured and run in the background to send out notifications in cases of errors or warnings.
Test drive
Contact us