I am currently documenting the performance of the Analytics Detail Query for organizations with more than ten thousand interactions per day. I am seeing significant latency when I try to query for more than twenty-four hours of data in a single request. Is there an undocumented limit on the ‘Page Size’ or the total number of ‘Intervals’ that can be included in a single analytics query before the performance starts to degrade?
Hello Cla41. I am an AI strategist and I have seen this latency cause issues with our Agent Assist training data exports. You should never query for more than thirty-one days of data in a single request. The API documentation says you can, but the reality is that the query will often time out or return a 504 error. It is much better to break your query into smaller, six-hour intervals and run them in parallel.
I use these analytics queries to find audio files for our training sets. I agree with Kaz85 about the intervals. Also, make sure you are using the ‘conversationEnd’ filter instead of just ‘interval’. It helps the analytics engine narrow down the search in the database much faster. I have seen a thirty percent improvement in response time just by adding that one filter!
I have been reporting these performance bugs to Genesys for months. The issue is that the analytics engine has to aggregate data from multiple shards for large queries. If you use too many ‘Dimensions’ or ‘Metrics’ in your query, it will be even slower. You should only request the absolute minimum data you need for your report. If you need everything, you should use the ‘Analytics Jobs’ API instead of the real-time query.