Process flow

The real-time runtime metrics collection process flow shows how the real-time runtime metrics collection data is processed and sent for data analysis.

Figure 1. Process flow for real-time runtime metrics collection
This figure shows an overview of the processing flow for real-time runtime metrics collection processing, which is described in the surrounding text.

  1. Real-time runtime metrics collection runs on the z/TPF system to collect data. Real-time runtime metrics collection runs continuously based on the time intervals that are defined by using the ZRTMC command. z/TPF system monitoring for Java™ applications determines the intervals that JVM data is collected and sends the data to real-time runtime metrics collection. For more information, see Data collection intervals.

    You can also use the ZRTMC command to start and stop real-time runtime metrics collection.

  2. The real-time runtime metrics collection data is sent in binary format from the z/TPF system to the tpfrtmc offline utility by using the high-speed connector. Because the high-speed connector is used, you can configure the z/TPF system to send data to multiple instances of the tpfrtmc offline utility, for example, to have a fallback tpfrtmc offline utility.
  3. The tpfrtmc offline utility processes the binary data from step 2. The tpfrtmc offline utility reduces the data and formats the results in a highly consumable JSON format. The processed data is sent to Apache Kafka.

    Start of changeA single instance of the tpfrtmc offline utility can process continuous data collection (CDC), system-wide JVM monitoring, and user-defined metrics data from multiple z/TPF processors simultaneously. For more information, see the maxconn field in the cdc, jvm, and udm section fields in the Runtime metrics collection properties file.End of change

    Start of changeA single instance of the tpfrtmc offline utility can accept name-value pair collection data from only a single z/TPF processor.End of change

  4. Apache Kafka functions as the communication mechanism of the data for your database, monitors, and analysis package. You can configure the Apache Kafka topics for different types of data by using the runtime metrics collection properties file.
  5. Connect your tools to Apache Kafka to display, analyze, or preserve the data.

For a sample analytics pipeline, see z/TPF real-time insights dashboard starter kit.