How to optimize costs in Azure Database for PostgreSQL - Flexible Server
APPLIES TO:
Azure Database for PostgreSQL - Flexible Server
Azure Database for PostgreSQL flexible server is a relational database service in the Microsoft cloud based on the PostgreSQL Community Edition.. It's a fully managed database as a service offering that can handle mission-critical workloads with predictable performance and dynamic scalability.
This article provides a list of recommendations for optimizing Azure Database for PostgreSQL flexible server cost. The list includes design considerations, a configuration checklist, and recommended database settings to help you optimize your workload.
- Leverage reserved capacity pricing.
- Scale compute Up/Down.
- Using Azure advisor recommendations.
- Evaluate HA (high availability) and DR (disaster recovery) requirements.
- Consolidate databases and servers.
- Place test servers in cost-efficient geo-regions.
- Starting and Stopping servers.
- Archive old data for cold storage.
1. Use reserved capacity pricing
Azure Postgres reserved capacity pricing allows committing to a specific capacity for 1-3 years, saving costs for customers using Azure Database for PostgreSQL flexible server. The cost savings compared to pay-as-you-go pricing can be significant, depending on the amount of capacity reserved and the length of the term. Customers can purchase reserved capacity in increments of vCores and storage. Reserved capacity can cover costs for Azure Database for PostgreSQL flexible server instances in the same region, applied to the customer's Azure subscription. Reserved pricing for Azure Database for PostgreSQL flexible server offers cost savings up to 40% for 1 year and up to 60% for 3-year commitments, for customers who reserve capacity. For more details, see Pricing Calculator | Microsoft Azure. To learn more, see What are Azure Reservations?
2. Scale compute up/down
Scaling up or down the resources of an Azure Database for PostgreSQL flexible server instance can help you optimize costs. Adjust the vCores and storage as needed to only pay for necessary resources. Scaling can be done through the Azure portal, Azure CLI, or Azure REST API. Scaling compute resources up or down can be done at any time and requires server restart. It's good practice to monitor your database usage patterns and adjust the resources accordingly to optimize costs and ensure performance. For more details, see Compute and Storage options in Azure Database for PostgreSQL flexible server.
Configure Non-prod environments conservatively - Configure idle dev/test/stage environments to have cost-efficient SKUs. Choosing Burstable SKUs is ideal for workloads that don't need continuous full capacity.
To learn more, see Scale operations in Azure Database for PostgreSQL flexible server
3. Using Azure advisor recommendations
Azure Advisor is a free service that provides recommendations to help optimize your Azure resources. It analyzes your resource configuration and usage patterns and provides recommendations on how to improve the performance, security, high availability, and cost-effectiveness of your Azure resources. The recommendations cover various Azure services including compute, storage, networking, and databases.
For Azure Database for PostgreSQL, Azure Advisor can provide recommendations on how to improve the performance, availability, and cost-effectiveness of your database. For example, it can suggest scaling the database up or down, using read-replicas to offload read-intensive workloads, or switching to reserved capacity pricing to reduce costs. Azure Advisor can also recommend security best practices, such as enabling encryption at rest, or enabling network security rules to limit incoming traffic to the database.
You can access the recommendations provided by Azure Advisor through the Azure portal, where you can view and implement the recommendations with just a few clicks. Implementing Azure Advisor recommendations can help you optimize your Azure resources and reduce costs. To learn more, see Azure Advisor for Azure Database for PostgreSQL - Flexible Server
4. Evaluate HA (high availability) and DR (disaster recovery) requirements
Azure Database for PostgreSQL flexible server has built-in node and storage resiliency at no extra cost to you. Node resiliency allows your Azure Database for PostgreSQL flexible server instance to automatically failover to a healthy VM with no data loss (that is, RPO zero) and with no connection string changes except that your application must reconnect. Similarly, the data and transaction logs are stored in three synchronous copies, and it automatically detects storage corruption and takes the corrective action. For most Dev/Test workloads, and for many production workloads, this configuration should suffice.
If your workload requires AZ resiliency and lower RTO, you can enable High Availability (HA) with in-zone or cross-AZ standby. This doubles your deployment costs, but it also provides a higher SLA. To achieve geo-resiliency for your application, you can set up GeoBackup for a lower cost but with a higher RTO. Alternatively, you can set up GeoReadReplica for double the cost, which offers an RTO in minutes if there was a geo-disaster.
Key take away is to evaluate the requirement of your full application stack and then choose the right configuration for the Azure Database for PostgreSQL flexible server instance. For example, if your application isn't AZ resilient, there's nothing to be gained by configuring Azure Database for PostgreSQL flexible server in AZ resilient configuration.
To learn more, see High availability architecture in Flexible Server
5. Consolidate databases and servers
Consolidating databases can be a cost-saving strategy for Azure Database for PostgreSQL flexible server. Consolidating multiple databases into a single Azure Database for PostgreSQL flexible server instance can reduce the number of instances and overall cost of running Azure Database for PostgreSQL flexible server. Follow these steps to consolidate your databases and save costs:
- Access your server: Identify the server that can be consolidated, considering database's size, geo-region, configuration (CPU, memory, IOPS), performance requirements, workload type and data consistency needs.
- Create a new Azure Database for PostgreSQL flexible server instance: Create a new Azure Database for PostgreSQL flexible server instance with enough vCPUs, memory, and storage to support the consolidated databases.
- Reuse an existing Azure Database for PostgreSQL flexible server instance: In case you already have an existing server, make sure it has enough vCPUs, memory, and storage to support the consolidated databases.
- Migrate the databases: Migrate the databases to the new Azure Database for PostgreSQL flexible server instance. You can use tools such as pg_dump and pg_restore to export and import databases.
- Monitor performance: Monitor the performance of the consolidated Azure Database for PostgreSQL flexible server instance and adjust the resources as needed to ensure optimal performance.
Consolidating databases can help you save costs by reducing the number of Azure Database for PostgreSQL flexible server instances you need to run and by enabling you to use larger instances that are more cost-effective than smaller instances. It is important to evaluate the impact of consolidation on your databases' performance and ensure that the consolidated Azure Database for PostgreSQL flexible server instance is appropriately sized to meet all database needs.
To learn more, see Improve the performance of Azure applications by using Azure Advisor
6. Place test servers in cost-efficient geo-regions
Creating a test server in a cost-efficient Azure region can be a cost-saving strategy for Azure Database for PostgreSQL flexible server. By creating a test server in a region with lower cost of computing resources, you can reduce the cost of running your test server and minimize the cost of running Azure Database for PostgreSQL flexible server. Here are a few steps to help you create a test server in a cost-efficient Azure region:
- Identify a cost-efficient region: Identify an Azure region with lower cost of computing resources.
- Create a new Azure Database for PostgreSQL flexible server instance: Create a new Azure Database for PostgreSQL flexible server instance in the cost-efficient region with the right configuration for your test environment.
- Migrate test data: Migrate the test data to the new Azure Database for PostgreSQL flexible server instance. You can use tools such as pg_dump and pg_restore to export and import databases.
- Monitor performance: Monitor the performance of the test server and adjust the resources as needed to ensure optimal performance.
By creating a test server in a cost-efficient Azure region, you can reduce the cost of running your test server and minimize the cost of running Azure Database for PostgreSQL flexible server. It's important to evaluate the impact of the region on your test server's performance and your organization's specific regional requirements. This ensures that network latency and data transfer costs are acceptable for your use case.
To learn more, see Azure regions
7. Starting and stopping servers
Starting and stopping servers can be a cost-saving strategy for Azure Database for PostgreSQL flexible server. By only running the server when you need it, you can reduce the cost of running Azure Database for PostgreSQL flexible server. Here are a few steps to help you start and stop servers and save costs:
- Identify the server: Identify the Azure Database for PostgreSQL flexible server instance that you want to start and stop.
- Start the server: Start the Azure Database for PostgreSQL flexible server instance when you need it. You can start the server using the Azure portal, Azure CLI, or Azure REST API.
- Stop the server: Stop the Azure Database for PostgreSQL flexible server instance when you don't need it. You can stop the server using the Azure portal, Azure CLI, or Azure REST API.
- Also, if a server has been in a stopped (or idle) state for several continuous weeks, you can consider dropping the server after the required due diligence.
By starting and stopping the server as needed, you can reduce the cost of running Azure Database for PostgreSQL flexible server. To ensure smooth database performance, it is crucial to evaluate the impact of starting and stopping the server and have a reliable process in place for these actions as required. To learn more, see Stop/start an Azure Database for PostgreSQL - Flexible Server instance.
8. Archive old data for cold storage
Archiving infrequently accessed data to Azure archive store (while still keeping access) can help reduce costs. Export data from Azure Database for PostgreSQL flexible server to Azure Archived Storage and store it in a lower-cost storage tier.
- Set up Azure Blob Storage account and create a container for your database backups.
- Use
pg_dump
to export the old data to a file. - Use the Azure CLI or PowerShell to upload the exported file to your Blob Storage container.
- Set up a retention policy on the Blob Storage container to automatically delete old backups.
- Modify the backup script to export the old data to Blob Storage instead of local storage.
- Test the backup and restore process to ensure that the archived data can be restored if needed.
You can also use Azure Data Factory to automate this process.
To learn more, see Migrate your Azure Database for PostgreSQL flexible server database by using dump and restore
Tradeoffs for cost
As you design your application database on Azure Database for PostgreSQL flexible server, consider tradeoffs between cost optimization and other aspects of the design, such as security, scalability, resilience, and operability.
Cost vs reliability
Cost has a direct correlation with reliability.
Cost vs performance efficiency
Boosting performance will lead to higher cost.
Cost vs security
Increasing security of the workload will increase cost.
Cost vs operational excellence
Investing in systems monitoring and automation might increase the cost initially but over time will reduce cost.
Next steps
To learn more about cost optimization, see: