PostgreSQL has firmly established itself as a leading database solution for workloads operating on Kubernetes, as highlighted in the Kubernetes in the Wild 2025 report. The report reveals that PostgreSQL now supports 36% of all database workloads on Kubernetes, marking a significant increase of 6 percentage points since 2022. This trend reflects the growing trust and preference for PostgreSQL within the Kubernetes ecosystem. However, managing data-intensive PostgreSQL workloads on Kubernetes presents its own challenges, including the complexities of Kubernetes primitives like StatefulSets and deployments, as well as the need for optimal performance through careful configuration of storage, replication, and database settings. Fortunately, advancements are simplifying this experience.
To cater to varying performance requirements, two distinct options for deploying stateful PostgreSQL workloads are now available. For databases that demand stringent latency and scalable transaction capabilities, Azure Container Storage can be utilized to orchestrate Kubernetes volume deployment on local NVMe, thereby enhancing IOPS while ensuring remarkably low sub-millisecond latency. In contrast, when price-performance optimization is paramount, Premium SSD v2 emerges as the preferred choice. Additionally, in collaboration with CloudNativePG, a robust open-source operator for PostgreSQL has been integrated to facilitate high availability database deployment on Azure Kubernetes Service (AKS). This combination of advanced storage options and CloudNativePG positions AKS as a formidable platform for high-performance PostgreSQL workloads.
Breakthrough PostgreSQL performance with local NVMe
For performance-critical PostgreSQL workloads, particularly those that manage vast concurrent transactions or require low-latency data access, local NVMe directly attached to Azure Virtual Machine (VM) SKUs stands out as an optimal solution. Historically, utilizing local NVMe drives with Kubernetes posed challenges, often necessitating RAID configurations and manual management of static volume orchestrators. Azure Container Storage effectively mitigates these complexities.
Azure Container Storage is a fully managed, container-native storage solution tailored for Kubernetes environments. Developers can seamlessly request a Kubernetes volume, prompting Azure to dynamically provision storage backed by the available local NVMe drives on AKS nodes. This setup provides PostgreSQL users with direct-attached block storage IOPS and latency within a managed, orchestrated cloud environment. Whether powering payment systems, gaming backends, or real-time personalization engines, users benefit from both speed and simplicity. Furthermore, Azure Container Storage supports Azure Disk and Elastic SAN (Preview), allowing for a choice of backing storage based on durability, scale, or cost as needs evolve—all under a consistent, Kubernetes-native control plane.
Benchmark results indicate that PostgreSQL can achieve nearly 15,000 transactions per second (TPS) with single-digit millisecond end-to-end query latency using the Standard_L16s_v3 VM. Scaling up to larger VM SKUs, such as Standard_L64s_v3, reveals TPS figures soaring to 26,000 while maintaining low latency. For further insights into these benchmark runs, please refer to the comparison of storage options section below.
Optimize price-performance with Premium SSD v2
Azure Premium SSD v2 strikes an ideal balance between price and performance, offering a flexible deployment model that suits production environments requiring scalability over time. With Premium SSD v2, users can independently configure IOPS, throughput, and size, enabling PostgreSQL deployments to dynamically scale with demand while minimizing upfront costs and avoiding resource overprovisioning.
This flexibility translates into real savings without compromising performance, particularly for multi-tenant SaaS platforms, production systems that grow alongside business needs, or applications experiencing fluctuating traffic. Premium SSD v2 supports up to 80,000 IOPS and 1,200 MB/s per volume, catering to highly demanding PostgreSQL workloads on an adaptable infrastructure.
Comparison of storage options
To assist in evaluating the aforementioned storage options, benchmark runs were conducted using CloudNativePG operator setups on AKS, maintaining similar core and memory consumption while varying only the backing storage: one utilizing local NVMe with Azure Container Storage and the other employing Premium SSD v2 with Disk CSI driver.
In the first configuration, a Standard_D16d_v5 SKU was used, provisioning two Premium SSD v2 32 GiB disks, each with 3000 IOPS and 125 MB/s throughput for log and data files. The second setup operated on Standard_L16s_v3 nodes with local NVMe storage included. The test environment was meticulously configured to replicate a real-world production database scenario. TPS measures the number of individual transactions (INSERT, UPDATE, DELETE, or SELECT) a system can handle per second, while latency denotes the delay between issuing a request to the database and receiving a response—crucial for applications requiring real-time or near-real-time responsiveness, such as financial systems, online gaming, or high-performance analytics.
Local NVMe on Standard_L16s_v3 achieved 14,812 TPS with an average latency of 4.321 milliseconds. In contrast, PremiumV2_LRS on Standard_D16ds_v5 recorded 8,600 TPS at 7.417 milliseconds latency. The pricing comparison is detailed below:
*Monthly costs are based on the base 3000 IOPS and 125 MB/s throughput. You can adjust the performance (capacity, throughput, and IOPS) of Premium SSD v2 disks at any time, allowing workloads to be cost-efficient while meeting size and performance requirements.
**With 3 VMs of L16s_v3, you receive 11.52 TB of storage allocated by default, utilized for the volumes created for PostgreSQL workloads. Pricing per month and allocated storage will vary for other VM sizes in the L-Series family.
When deciding between local NVMe and Premium SSD v2 for PostgreSQL workloads, considerations include balancing performance, cost, and data durability. Local NVMe via Azure Container Storage offers extremely low latency and high throughput, making it ideal for performance-sensitive PostgreSQL deployments. However, costs may be higher, and scaling flexibility is limited. Conversely, Premium SSD v2 provides superior price-performance efficiency and scalability, making it suitable for PostgreSQL instances that must accommodate increased scale or unpredictable demand surges. In terms of data durability, Premium SSD v2 offers local redundancy by default, while local NVMe is best paired with a replica-based architecture managed by CloudNativePG operator and an object storage-based backup approach to mitigate data loss.
Built for high availability with CloudNativePG on Azure Kubernetes Service
For teams deploying PostgreSQL in production environments, high availability and backups are essential. The open-source CloudNativePG operator simplifies the deployment of a highly available PostgreSQL cluster on AKS, featuring:
- Built-in replication and automated failover.
- Application consistent backup with native integration with Azure Blob Storage.
- Seamless integration with Azure Container Storage.
- Flexible storage options: choose Premium SSD v2 or local NVMe based on workload needs.
This framework provides peace of mind for organizations, eliminating the need for custom high availability logic and separate backup workflows. To begin deploying a highly available PostgreSQL environment on AKS with CloudNativePG operator, refer to our step-by-step reference guide.
Ready for the future
PostgreSQL is among the many stateful workloads that organizations are now confidently deploying on Azure Kubernetes Service. From databases to message queues, AI inferencing, and enterprise applications, AKS is evolving to support the demands of persistent, data-intensive applications in production.
Whether deploying Redis, MongoDB, Kafka, or even ML-serving pipelines with GPU-backed nodes, AKS provides a solid foundation for managing these workloads with performance, consistency, and operational ease, complemented by clear end-to-end guidance.
With innovations such as Azure Container Storage for local NVMe and Premium SSD v2 for scalable persistent storage, building stateful applications that are reliable, performant, and cost-efficient for mission-critical workloads has never been easier.
Modernize your data layer on Kubernetes today. Whether running PostgreSQL or any stateful tier, Azure delivers the performance and manageability necessary to achieve your goals. Explore proven patterns and deployment options in the AKS Stateful Workloads Overview.