When using Document API responses for monitoring what is a reasonable response time before triggering an alert? - c06330146

Some customers use Document API calls to measure of the health of the service. When doing this, what is a recommended response time and methodology for using DocAPI for health monitoring?

When using the Document API for measuring the status of a service, the tests should consider that rendering times can vary from document to document and also for specific drivers. Typically a full print completion process is not used to measure health checks since the results can be highly variable.

There are 2 approaches to tracking server health with the Document API

1. Use DocStatus - In this case the client should perform a DocStatus call - with the understanding it will not be successful. The purpose is to validate the response by locating "jobState" in the XML response. Expected Response times should be less than 1 second
2. Full DocAPI Print - Perform a full DocAPI submission and status check. Submit a known good document to a known valid printer using DocProcess. Track the job status using DocStatus until the jobState reaches a value of 9 - success

NOTE:

The time to wait should be determined by the test client prior to creating the test.

Send a test job multiple times to determine the average print duration. Add some buffer time (such as 25% - 50%) depending on the typical server load