Azure SQL Database Pricing
SQL Database billing has three components — compute, storage, and backup storage:
| Component | Billing Unit | Approx. Cost (Central India) |
|---|---|---|
| Compute (vCore) | Per vCore per hour | General Purpose: ~₹6–8/vCore/hour |
| Storage | Per GB per month | ~₹10–12/GB/month |
| Backup storage | Per GB per month above free | First 100% of DB size free; LRS ~₹1/GB beyond |
Real Example — General Purpose, 4 vCores, 100 GB
| Component | Monthly Cost |
|---|---|
| 4 vCores × ~₹7 × 730 hours | ~₹20,440 |
| 100 GB storage × ₹11 | ~₹1,100 |
| Backup (within free tier) | ₹0 |
| Total | ~₹21,540/month |
SQL Managed Instance Pricing
SQL MI is significantly more expensive than SQL Database — it provides a full SQL Server instance:
| Tier | vCores | Approx. Monthly Cost |
|---|---|---|
| General Purpose | 4 vCores | ~₹50,000–60,000/month |
| General Purpose | 8 vCores | ~₹100,000–120,000/month |
| Business Critical | 4 vCores | ~₹160,000–180,000/month |
Cosmos DB Pricing
Cosmos DB bills on RU/s (throughput) and storage:
| Component | Cost |
|---|---|
| Provisioned throughput (manual) | ~₹6 per 100 RU/s per hour |
| Autoscale throughput | ~₹7.5 per 100 RU/s per hour (billed at current scale) |
| Serverless | ~₹0.024 per 10,000 RUs consumed |
| Storage (transactional) | ~₹21/GB/month |
| Additional write regions | Each additional region multiplies RU/s cost |
MySQL & PostgreSQL Pricing
Both MySQL and PostgreSQL Flexible Server share a similar pricing model:
| Tier | vCores | Approx. Monthly Cost (Central India) |
|---|---|---|
| Burstable | 1 vCore | ~₹1,200–1,800/month |
| Burstable | 2 vCores | ~₹2,500–3,500/month |
| General Purpose | 2 vCores | ~₹8,000–10,000/month |
| General Purpose | 4 vCores | ~₹16,000–20,000/month |
| Business Critical / Memory Optimised | 2 vCores | ~₹16,000–20,000/month |
Stop/Start Saves Money
MySQL and PostgreSQL Flexible Servers can be stopped when not in use. Compute billing pauses — storage billing continues (~₹10/GB/month). A dev server stopped for 16 hours per day saves ~67% on compute.
Redis Cache Pricing
| Tier | Cache Size | Approx. Monthly Cost |
|---|---|---|
| Basic C0 | 250 MB | ~₹1,000/month |
| Standard C1 | 1 GB | ~₹3,500/month |
| Standard C2 | 6 GB | ~₹9,000/month |
| Premium P1 | 6 GB | ~₹18,000/month |
| Premium P3 | 26 GB | ~₹75,000/month |
Synapse Analytics Pricing
| Component | Cost |
|---|---|
| Dedicated SQL Pool (DW100c) | ~₹55/hour → ~₹40,000/month (if always on) |
| Serverless SQL Pool | ~₹370/TB scanned |
| Spark Pool | Per vCore-hour when running |
| Data Lake Storage | Standard storage rates |
Azure Hybrid Benefit
If you have existing SQL Server licences with Software Assurance (SA), Azure Hybrid Benefit (AHB) allows you to use them in Azure — reducing costs by 30–40%.
| Service | AHB Discount |
|---|---|
| Azure SQL Database | Up to 30% off licence component |
| SQL Managed Instance | Up to 40% off licence component |
| SQL Server on VM | Use own licence — save full SQL Server licence cost |
Reserved Capacity
For predictable, steady-state workloads, reserve capacity 1 or 3 years in advance for significant discounts:
| Service | 1-Year Reservation | 3-Year Reservation |
|---|---|---|
| Azure SQL Database | ~33% discount | ~60% discount |
| SQL Managed Instance | ~33% discount | ~60% discount |
| Cosmos DB (RU/s) | ~17% discount | ~29% discount |
| MySQL/PostgreSQL | ~33% discount | ~60% discount |
| Redis Cache | ~33% discount | ~60% discount |
Top 10 Database Cost-Saving Tips
1. Use Azure Hybrid Benefit
If you have SQL Server licences with SA, enable AHB before deploying — saves 30–40% immediately.
2. Buy Reserved Capacity for Stable Workloads
3-year reserved capacity saves up to 60% vs pay-as-you-go for always-on databases.
3. Use Serverless for Dev/Test
SQL Database Serverless auto-pauses — billing stops for compute when idle.
4. Stop MySQL/PostgreSQL Servers After Hours
Flexible Servers can be stopped manually or via Azure Automation. Stop dev servers overnight.
5. Pause Synapse Dedicated SQL Pools
Biggest Synapse cost lever — pause when not running analytics queries.
6. Right-Size Before Deploying
Start with the smallest tier that meets requirements. Scale up is easy — but scale down requires careful planning.
7. Use Elastic Pools for Multi-Tenant Apps
Multiple databases sharing a pool costs far less than individual databases for SaaS apps with variable usage.
8. Choose the Right Cosmos DB Throughput Mode
Serverless for dev/test. Autoscale for variable production. Manual provisioned for predictable high-volume.
9. Minimise Cosmos DB Write Regions
Each write region multiplies RU/s cost. Use read-only regions where possible — much cheaper.
10. Monitor with Azure Cost Management
Set budgets and alerts. Review cost analysis monthly. Identify orphaned databases and unused elastic pool capacity.