Logging for Nuance speech products

Log files contain errors, events, and application transactions in chronological order. During a session, all Nuance components share a unique session ID, which you can use to map all the messages, alarms, and events for that session.

Speech Server creates a single unified call log for each session (call), including events from Nuance speech products. The logging system consists of a call log server, which is part of Speech Server and installed with it, and one or more call log clients, which are part of the respective speech products. During a telephone call, the call log server collects and merges data. At the end of a session, the call log server saves the data in the root directory designated by the server.callLog.baseDirectory parameter on the Speech Server, and subdirectories determined by the company_name and application_name parameters in session.xml. The data includes text log files, audio utterances, whole call recording audio files, wave forms, diagnostic logs, and so on.

Types of logs:

• Call logs: Text files showing minute detail about every call that connects to the system. Useful for evaluating and tuning application performance. Nuance speech products write many application events automatically (see the standard events in Tokens and events in call logs).
• Diagnostic logs: Detailed record of system activity, communications, and errors. Useful for troubleshooting applications and installations.
• Audio files: Waveforms containing the utterances spoken by callers. Useful for transcription and tuning.

You can configure logging for different phases in your development cycle. The logs are useful to different users at different times in the lifecycle:

• System administrators determine the logging configuration and root location for storing log files. During installation, system administrators configure the logging service for properties such as size, caching, log file cleanup, and so on. They can use default settings or they provision the system for multiple applications and companies.

  Administrators also monitor system performance, and evaluate the capacity of the speech system. They use diagnostic logs when investigating system latency.

• Application developers set the level of detail logged for events, decisions, and transactions during application development, deployment, and tuning. For example, you may log more information during development, and only FATAL events during full deployment.

  • During development, they design call log messages to record significant application events, decisions, and transactions.

  • During deployment, application developers provide for runtime settings, such as the company and application names written to the logs. Developers review logged information during initial deployment, make tuning adjustments, and monitor the results. They also use diagnostic logs throughout the development and deployment process.

• Platform integrators who access Nuance speech products from their voice browsers use diagnostic logs to troubleshoot their integrations. (See Diagnostic logging

Nuance Speech Suite automatically protects data in logs. Application developers and system administrators can control additional features to suppress or encrypt information when needing strict data protection. The system can protect any confidential data. Typically, this means personal data such as names, addresses, telephone numbers, account numbers, passwords, and so on. You can configure the level of security and specify which interactions involve confidential information. For details on protecting confidential data, see Protecting confidential data.