A ticket report needs a defined population before its count can mean anything. A created cohort selects tickets by creation time. A solved cohort selects records using the supplied solved date. They answer different questions even when the date range is identical. Current backlog means unresolved work at the current training time, not a historical backlog reconstruction. Keep one row per ticket and name exclusions such as archived or closed records where the report model applies them.
Build and run the requested report, then compare the output with its contributing tickets. Later tasks alter the cohort or measure so a familiar total is insufficient. These are original fictional support records. The local model derives selected current measures and does not reproduce a full analytics warehouse.
Worked example: ticket cohorts and backlog.