QV: Monitoring patterns for common problems

    This document summarizes the list of log patterns that are generated when QVS (11SR12, IIS) is down or unavailable for users in my experience.

     

    QlikView event log entries at level 300 (severe) are not always produced when the instances are down/restarting or hung.

    I am hoping to grow a comprehensive  list of entries that can be monitored by OS system management tools and used for alerting administrators, without having to set up elaborate log processing procedures.

     

    QV Event Log entries:

    1. 'NUMA DETECTED' -- produced every time QVS service restarts or recovers from a crash.(could be CPU architecture dependent)

    2. 'Phoenix detected' -- produced when QVS hang is detected by internal monitoring thread.

     

    Web Server Log entries(when debug/verbose logging is enabled):

    1. 'Information LoadBalancing - selected Qvs: None' -- produced when "No Server" error message is displayed on Access Point for user.

    2. 'The communication with QlikView Server failed' -- produced when active session is terminated by QV server crash.

     

    Please feel free to contribute.