1. Performance
1.1. Esky
1.1.1. Available dates for segments (metrics_index)
1.1.2. Rolling data for Weekly, 2 Weeks, and so on.
1.1.3. Baked in TopX column into Esky schema
1.1.4. Aggregate Industry Tables for Search and Clickstreams tables
1.2. Tableau Workbooks
1.2.1. Configure GP connection setings
1.2.1.1. Disable validation queries.
1.2.1.2. Increase cursor size when fetching huge data
1.2.2. Pre-calculated fields in GP.
1.2.2.1. Industry Top Terms extracts
1.2.3. Reduce data volume in extracts
1.2.3.1. Increase thresholds in query
1.2.3.1.1. Industry Top Terms extract
1.2.3.1.2. Custom Industry Top Search Terms extract
1.2.3.2. Reduce date range to latest single week for non trending chart
1.2.3.2.1. Custom Industry Up/Downstream Websites extract
1.2.3.2.2. Custom Industry Top Search Terms extract
1.2.4. Avoid having more than 1 query being executed per worksheet.
1.2.4.1. Change from using published Datasource to local.
1.2.5. Document performance Do's and Don'ts
2. Development
2.1. Refactor workbooks to make it simpler
2.1.1. Removed unnecessary extra worksheet such as table header in Control Panel comparison report
2.1.2. Use URL Action to escape multi-value separator instead of implementing custom field
2.1.3. Remove duplicate parameters
3. Challenges
3.1. Tableau
3.1.1. Control Panel
3.1.1.1. Too complex to achieve wireframe
3.1.1.2. Ajax load is very slow at first load
4. Stability
4.1. Tableau
4.1.1. Extract out the data loading process out of Tableau Extract refresh
4.1.1.1. script + Jenkins
4.1.1.2. Pentaho data integration (PDI)
4.1.2. Random Exceptions by Tableau Server
4.2. Esky
4.2.1. Out of Memory
4.2.1.1. Increase memory setting (GC)
4.2.2. Timestamps to help extract refresh incrementally
4.2.3. Slower during builds
5. Monitoring
5.1. Splunk
5.2. Tableau extracts history
5.3. Server health checks
6. Deployment
6.1. Continuous deployment
7. QA
7.1. Esky
7.1.1. Data integrity
7.1.2. Automation
7.2. Tableau Server
7.2.1. Automation