All challenges
Large Data VolumesArchitect
Taming a 200M-Row Object
3 decision points685 grounded questions
The scenario
Transaction__c has 200 million rows and grows 5M/month. List views and reports time out, an after-insert trigger that aggregates onto the parent Account now throws 'non-selective query against large object type', and the org is nearing its data storage allocation. You must restore performance and control cost without losing data the business still queries.
Constraints to satisfy
- Reporting on the last 18 months must stay fast.
- The aggregation trigger must not throw on large operations.
- Older-than-18-month data must remain queryable but cheaply stored.
- No third-party data warehouse is approved yet.
Think through your own design first - data model, timing, failure modes. Then walk the decision points and defend each choice against a worked architecture.