Information:

Environment

The functionality is available from AppMon version 6.5. The target system must be 6.5 (or higher), the source system 6.1 or higher for measurement data and percentiles. If also incidents should be imported, the source system must be at least version 6.3.

Icon

This functionality is currently BETA, so it is not fully supported and you are using it on your own risk! Therefore, a PWH backup is recommended.

Symptoms

Use the tool to migrate measurement data from different Dynatrace AppMon servers or to merge or split system profile measurements.

Solution
 As this functionality is beta in AppMon 6.5, handle with care!
  1. Export the system profile from the source system and import it on the target system. Alternatively, you can import the data also to an already existing system profile. For all measures with identical names and definitions the data will be imported.
  2. Set the client in debug mode (Ctrl+Shift+F9)
  3. Open the "Tasks and Monitors" dashboard and create a new task for the system profile into which you would like to import the data.
  4. Select the "System profile data import" task
  5. Configure the correct settings
    1. Set the data for the source database where the data is stored at the moment
    2. Set the name of the source system profile
    3. Activate the "measure splittings importer" checkbox. This will import the dynamic measures (splittings of a measure)
    4. If you want to import measurements, activate "measurements importer"
      1. Set the number of days of low resolution data you would like to import
      2. Set the number of days being imported in one round trip. Note: high value: faster performance, higher memory usage; low value: slower performance, less memory usage
      3. Set the duration of hours of mid resolution data you would like to import
      4. Set the number of hours being imported in one round trip.
      5. Set the duration of minutes of high resolution data you would like to import. Note: it isn't recommended to import high resolution data, as this needs a high number of resources and the data will be aged out relatively fast.
      6. Set the number of minutes being imported in one round trip.
    5. If you want to import incidents, activate "incidents importer"
      1. Set the duration of hours of incidents data you would like to import
      2. Set the number of hours being imported in one round trip. Note: high value: faster performance, higher memory usage; low value: slower performance, less memory usage
    6. If you want to import percentiles, activate "percentiles importer"
      1. Set the duration of hours of percentiles data you would like to import. Note: as percentiles data are aged out with measurement low measurements, this can be a very large amount of data
      2. Set the number of hours being imported in one round trip. Note: high value: faster performance, higher memory usage; low value: slower performance, less memory usage
    7. Set the "max execution time for the task". The "System profile data import" task is intended to run for some hours each day, until the import of the data has been finished. Therefore, set the hours the task should run each day.
    8. Set the "reference timestamp", if needed. This time stamp can be used as offset, if the data import should not start from the moment the task is defined.
    9. Test the database connection to make sure that the data importer task can access the source database
  6. Schedule the task. Decide if you would like to start it manually or if it should run on a daily basis.
  7. Press ok and start the task manually or wait for it to be executed if you have set a specific time to start

The status of the data import can be observed with the incidents "Performance Warehouse Import" and "Performance Warehouse Import Ongoing" in the target system profile. As both incidents are informational incidents, the dashlet filter options might have to be changed.

Example settings for the data importer task

Icon

Depending on the amount of data which should be imported, the task will need some time to finish. For an XLarge environment, a runtime of 27 days has been observed, most of the time needed for the import of the percentile data.