Database Capacity Planning: Storage Growth, Index Size, IOPS & Query Cost

What you'll build

What you will be able to do

Each lab ends with a concrete result that you build yourself and the checker confirms, cell by cell.

• Estimate how big Parcelwise's shipments table really is: bytes per row from the column types, rows that fit on an 8 KB page and the table size for 48 million rows.
• Size the four B-tree indexes on the shipments table: entries per page with a 90 % fill factor, leaf pages, MiB per index and how big all indexes are compared with the table itself.
• Project Parcelwise's database for the next 12 months, add the overhead the database needs on top of its data, find when the 80 % disk alert will fire and chart it.
• Turn Parcelwise's query mix into the disk numbers a storage volume is sold by: read IOPS, write IOPS with write amplification, the peak IOPS and the peak throughput in MiB/s.
• Measure how well Parcelwise's cache works, estimate the hot working set, size the buffer pool with headroom and pick the smallest instance with enough RAM.
• Add up everything Parcelwise stores for one 640 GB database: replicas, the full backups a 30-day point-in-time recovery window needs, the archived WAL, and the multiplier over the primary.
• Size the connection pools of Parcelwise's five services from their peak load, check them against max_connections and see how much load a CPU-based pool guideline can carry.
• Compare the monthly cost of four provisioned options with a serverless database that scales with Parcelwise's daily load, and chart the result.