Improving Process Logging Performance
To improve process logging performance, consider the following:
Choose
2 - Errors only,
3 - Process only, or
4 - Process and start steps as your logging level. Choose
5 - Process and all steps only when you need ultimate quality of service.
Store input pipelines only when absolutely necessary. It is usually sufficient to store pipelines for failed steps only. Remove all unnecessary data from pipelines to minimize the volume of data to store.
For process steps that run services, there are two areas in which you could inadvertently log the same information twice:
Process Engines can write start and successful completion or failure log entries for process steps that run services. Services can write log entries that convey the same information.
Process Engines can store input pipelines for services that are run by process steps. Services can also log input pipelines.
Coordinate your logging for these services to avoid logging the same information twice.
Note: When coordinating logging, consider that when a service is run by a process step, that service is actually called by a wrapper service, making it a nested service (as opposed to a top-level service).
For instructions on setting up service logging, and for complete information on logging in general, see webMethods Audit Logging Guide.