Jump to: navigation, search

Advisors Genesys Adapter Performance Information

The following performance results were achieved with Advisors Genesys Adapter software release 8.5.1. To avoid AGA performance issues, both the statistics load and maximum message rate must be below the thresholds described in the following table. See Determining Message Rates and Estimating the Number of Requested Statistics for Frontline Advisor for more information.

AGA Performance Information
Advisors Application Configuration Message Rate Stats Load Notes and Recommendations
Contact Center Advisor One AGA and one Stat Server Maximum message rate under 70 000 messages/second

Average message rate under 18 000 messages/second

AGA can process up to 1 200 000 statistics without performance degradation.

To generate 1 200 000 statistics, Genesys used the following configuration during testing:

  • Three time profiles enabled
  • FA hierarchy: 11 320 agents
If your statistics load is more than 1 200 000, then add one more Stat Server.

If your maximum message rate exceeds 70 000 messages/second, then add one more AGA.

Contact Center Advisor One AGA and two Stat Servers Maximum message rate under 45 000 messages/second

Average message rate under 6500 messages/second

AGA can process up to 891 000 statistics without performance degradation.

To generate 891 000 statistics, Genesys used the following configuration during testing:

  • Agent monitoring: On
  • Number of queues: 550 (with 3 time profiles)
  • Number of agent groups: 25 396
If your statistics load is more than 891 000, then add one more AGA.

If your maximum message rate exceeds 45 000 messages/second, then add one more AGA.

Frontline Advisor One AGA and one Stat Server Maximum message rate under 70 000 messages/second

Average message rate under 18 000 messages/second

AGA can process up to 1 200 000 statistics without performance degradation.

To generate 1 200 000 statistics, Genesys used the following configuration during testing:

  • Three time profiles enabled
  • FA hierarchy: 11 320 agents
If your statistics load is more than 1 200 000, then add one more Stat Server.

If your maximum message rate exceeds 70 000 messages/second, then add one more AGA.

Frontline Advisor One AGA and two Stat Servers Maximum message rate under 70 000 messages/second

Average message rate under 18 000 messages/second

AGA can process up to 1 450 000 statistics without performance degradation.

To generate 1 450 000 statistics, Genesys used the following configuration during testing:

  • Three time profiles enabled
  • FA hierarchy: 13 680 agents
If your statistics load is more than 1 450 000, then add one more AGA.

If your maximum message rate exceeds 70 000 messages/second, then add one more AGA.

Determining Message Rates

It is important to know how many messages arrive per second at the AGA from Stat Server to ensure you have a sufficient number of adapters deployed to handle the load. Use the following procedure to determine the message rates for your AGA.


Procedure: Calculating the message rates


Steps

  1. Locate your log4j.properties file in the <AGA installation>\conf folder.
  2. Change the log4j.logger.timing log entry to:
    log4j.logger.timing=debug, timingLog
  3. Check for new entries in the timing.log file. You should see the following type of entries:
    ... DEBUG timing - For object type: XXXX received xxxx messages from SS:xxx ...
  4. Wait until AGA accumulates sufficient data, and then you can determine the maximum number of messages received per second from Stat Server. You can also calculate the average number of messages received per second from Stat Server.

Estimating the Number of Requested Statistics for Frontline Advisor

You can evaluate the number of statistics requested for each agent in Frontline Advisor. When you have an estimate of the number of requested statistics, you can also estimate the following:

  • the number of Stat Servers to deploy
  • the FA statistics load on each Stat Server

The number of statistics requested for Frontline Advisor depends on the following:

  • the number of time profiles currently enabled
  • the number of report metrics currently enabled, including all dependencies that are not enabled
  • the number of agents currently logged on

Initially, estimate the load based on the enabled default metrics, or based on the metrics that are enabled in the migrated environment. If you change the number of enabled metrics after Advisors installation or migration, or if you make changes to time profiles, then use the same process to re-evaluate the load.

If necessary, you can do a post-installation adjustment of the Stat Server configuration to achieve optimal performance for your enterprise. See the Genesys Pulse Advisors Deployment Guide for information about Advisors Stat Server configuration.


Procedure: Estimating the number of statistics requested for each agent in Frontline Advisor


Steps

  1. Use the following queries to determine the number of statistics requested for each agent in Frontline Advisor:
    SELECT COUNT ( * ) "Perf. metrics" from fa_vw_performance_source;
    SELECT COUNT ( * ) "Metrics for rules p/agent" from fa_vw_rule_source;
    SELECT COUNT ( * ) "State metrics p/agent" from fa_vw_state_source;
    The first query (SELECT COUNT ( * ) "Perf. metrics" from fa_vw_performance_source;) provides the number of performance metrics. The second query (SELECT COUNT ( * ) "Metrics for rules p/agent" from fa_vw_rule_source;) provides the number of rules metrics. The third query (SELECT COUNT ( * ) "State metrics p/agent" from fa_vw_state_source;) provides the number of state metrics.
  2. Use the following calculation to estimate the overall, real-time number of statistics that will be sent to the configured Stat Servers:
    <Number of statistics obtained from running the queries> * <Average number of agents typically logged on to the system>

Next Steps

You can use the following query to find the number of performance metrics, grouped by time profile (this shows you how many raw metrics will be requested from Stat Server):
  1. SELECT COUNT ( * ) "Perf. metrics p/tprofile",type "Time Profile Type",interval "Time Profile Interval" from fa_vw_performance_source group by type,interval;

Feedback

Comment on this article:

blog comments powered by Disqus
This page was last modified on February 1, 2018, at 21:24.