Data that keeps accumulating.
WHAT GROWS
Transactional records Logs and audit tables Sessions Soft-deleted rows Indexes, alongside the data
WHAT TO MEASURE
Size per table, and its growth rate.
HOW
Query the table metadata for data and index sizes.
WHAT TO PROJECT
When the working set exceeds memory When the disk fills
WHY THE FIRST MATTERS MORE
Performance degrades sharply once data no longer fits in memory.
WHAT TO DO ABOUT LARGE HISTORICAL TABLES
Archive old rows elsewhere.
WHAT ARCHIVING REQUIRES
Agreement on what may be moved A destination A process that does not lock the table
WHAT PARTITIONING PROVIDES
Splitting one table into parts, usually by date.
WHAT IT ENABLES
Dropping a whole period instantly Queries touching only relevant partitions
WHAT IT COSTS
Complexity, and constraints on keys.
WHAT TO DELETE IN BATCHES
Large deletions, never in one statement.
WHY
A single enormous delete holds locks and generates vast undo.
WHAT TO REVIEW
Whether old data must be retained at all.
WHAT TO CHECK BEFORE DELETING ANYTHING
Retention obligations.
