What Happens When Usage Grows

The database first, always

Business applications rarely become slow because PHP cannot keep up. They become slow because query count grows with data volume, or because an index that was unnecessary at ten thousand rows is essential at a million.

Adding application servers in front of a database that is the bottleneck makes the bottleneck worse, not better.

The order of work

1. Fix query multiplication — loops issuing one query per row
2. Add missing indexes on filtered and joined columns
3. Move reporting to a replica or a nightly copy
4. Cache what remains genuinely expensive
5. Then add capacity, if it is still needed

Watch data volume, not user count

• A query that scans a table gets slower as the table grows
• Reports over all history get slower every month
• Log and audit tables grow indefinitely unless managed
• Search over free text degrades without proper indexing

Applications frequently slow down with no change in usage, purely because the data has grown past what the original queries assumed.

Archive old data

Most business applications keep everything forever because nobody decided otherwise. Moving old records to an archive table or a separate store keeps the working set small and the queries fast.

It also makes backups faster and restores quicker, which matters more than it sounds during an incident.

Then horizontal capacity

Frequently asked questions