Checking Package/User Histories

Always begin analyzing application errors or issues by checking high-level application statistics. The goal is to see if there are any obvious outliers in typical or expected application performance. The more familiar an administrator is with the environment, the more easily the administrator can identify abnormal or unexpected behavior or performance.

1. In Sybase Control Center, click the Monitoring node, then click the tab that corresponds to the application type you are investigating, either Replication or Messaging.
2. Click History, then choose Summary, to display an aggregated history for the package.
3. Select Show current filter, to narrow the results. Using the troubleshooting data you gathered from the user, in the filter pane:
   1. Select the domain to which the package is deployed.
   2. Select the package name from the list of packages deployed to the domain.
      Note: In the Domain and Packages fields, you can start to type a package or domain name to narrow the list to names beginning with the characters you enter.
4. Click the User column to sort in alphabetical (ascending or descending) order.
5. Refine entries to those that fall within the documented time frame. Set the Start Date, Start Time and Finish Date, Finish Time to restrict the data to the required duration.
6. Determine whether the volumes are high for any of the cells in a row. If so, further investigate the details of that package/user pairing row.
7. On the History tab, choose Detail view and locate the package/user row you are investigating. This view details the synchronization request, so you can find out where in the transaction chain the problem is arising. It helps you to identify what happened at the entity level.
8. Look at these columns:
   • Total Rows Sent to see if the number of data rows being synchronized is high or low.
   • Payload size to see the number of bytes downloaded to the device for each event.
   These counters may indicate that there is a lot of data being transferred during synchronization and may require some intervention to improve performance. You can then look at the entity or phase where the problem is occurring, and whether or not there is a delay of any kind that might indicate the nature of the problem. For example, perhaps large volumes of uploaded data may be delaying the download phase. However, it may also be that this behavior is normal, and the performance lag transitory.