Azure Table Storage – Partition Strategy Simulator

Simulate how different PartitionKey strategies affect data distribution, throughput, and query performance in Azure Table Storage. Choose a strategy, configure your workload, and see the impact visually.

⚙️ Configuration

Partition Key Strategy

Total Entities 100,000

Time Span (days) Simulation parameter: 1 – 3,650 days (~10 years)

Number of Categories Simulation parameter: 1 – 1,000 categories

Target Operations/sec Azure limit: up to 20,000 ops/sec per storage account

Primary Query Pattern

Entity Arrival Distribution

📊 Results Summary

Total Partitions

Max Entities / Partition

Avg Entities / Partition

Peak Ops / Partition / sec

📈 Partition Distribution

Entity count per partition (top 50 shown if many partitions)

🌡️ Throughput Heatmap

Estimated operations/sec load per partition

💡 Recommendations

📚 How Partition Strategy Affects Performance

Azure Table Storage Partition Limits:
• Each partition can handle up to 2,000 entities/sec throughput
• Entity Group Transactions (batch) limited to 100 entities in same partition
• Single entity max size: 1 MB; single partition: effectively unlimited rows
• Table-level target: 20,000 ops/sec per storage account

Strategy	Best For	Watch Out
Date-based (day)	Time-series, logs with date queries	Today's partition = hot spot
Category-based	Few, well-defined groups; batch ops	Uneven category sizes → hot partitions
Unique (GUID)	Max parallelism, point queries	No batch ops, no range scans
Composite	Balances grouping with distribution	More complex key management
Hash-bucket	Write-heavy, even distribution needed	Loses natural ordering; fan-out reads