As a follow-up to #9138, we want to improve test coverage of template insights. This means we'll also need manually generate/insert agent statistics which are a basis for some of the data.

Test cases should include:

• Multiple templates
• Multiple users
• 1-2 apps in 1-2 templates
• Controlled app usage data, e.g. user X used template Y / app Z between V and W
• Controlled agent usage data (agent stats)

What to test:

• Template ID filtering conditions (none, 1, 2, etc.)
• Requests in different timezones
• Resulting numbers must match what's expected from crafted input data