About the Metering Simulator
Integration Server includes a metering simulator that gathers service metrics and associated transaction information and writes that information to the server log. When the metering simulator is enabled, Integration Server produces a server log message for each service that would be metered. Integration Server writes one log message per invocation of a top-level, user-defined service. Integration Server does not aggregate data across multiple invocations of the same service into a single log message.
The metering simulator (and the webMethods Metering Agent) generates metrics for top-level user-defined services only.
Following is an example of server log message for a metered service:
2021-05-20 16:55:35 EDT [ISS.0176.9999I] (tid=92) Simulated metering metrics for service: myFolder.mySubFolder:exampleService, tenant=null, cpu=0(ns), duration=0(ms), transactions=1
The format for the message is:
Simulated metering metrics for service:serviceName, tenant=ID, cpu=cputime(ns), duration=durtime(ms), transactions=count
Where:
serviceName: The name of the top-level service for which the simulator is generating metrics.
tenant: The tenant ID. For a self-hosted runtime, the tenant ID will be null.
cpu: The amount of CPU time used for the service, measured in nanoseconds.
duration: The duration of the service, measured in milliseconds.
transactions: The number of transactions counted for this single service invocation. If the duration is less than 3 seconds, the count will be 1. If the duration exceeds 3 seconds, additional transactions will be counted for each 3 second interval. If your license key defines a different transaction duration value, the simulator will use that value to determine the number of transactions.
Note:
The (tid) is the thread ID of the thread performing the service execution.
The metering simulator runs independently of webMethods Metering. The metering simulator only enables the logging of messages for services that would be metered and therefore generate transaction data. Enabling the metering simulator does NOT result in actual reporting of the measured transaction data nor does it result in charges.
The metering simulator has the following main use cases:
Your site has the traditional perpetual license and you are considering moving to the transaction-based pricing model. Enabling the metering simulator for a period of time, and then analyzing the resulting data may provide guidance on how many transactions per month your site would need. The period of time for which you enable the simulator should reflect a typical transaction load.
Your site already has a subscription, usage-based license but you want access to critical metering data that is not reported with
webMethods Metering. Specifically, the metering simulator provides the names of the top-level services that generate the transactions and the number of transactions resulting from each top-level service execution. The logs generated by metering simulator can be key to understanding your usage data. For example, if you see a
webMethods Metering report that indicates your site ran a million transactions last month, you may want to know the source of those transactions: which
Integration Server generated the most transactions and which services are long-running and contributing the most transactions. Using the server logs created while the metering simulator is running can help you determine the answer to these questions.