This page describes the approximate performance that Spanner can provide under optimal conditions, factors that can affect performance, and tips for testing and troubleshooting Spanner performance issues.
The information on this page applies to both GoogleSQL and PostgreSQL databases.
Performance and storage improvements
Performance and storage improvements have rolled out to all Spanner regional, dual-region, and multi-region instance configurations. You don't need to make any changes to your application or manually configure anything in your Spanner instances to take advantage of these improvements, which are offered at no additional cost. These performance improvements result in higher throughput and better latency in Spanner nodes in all instance configurations.
Increased performance throughput
All Spanner instance configurations have improved performance and offer increased throughput. The following table provides the approximate throughput (queries per second) for Spanner instance configurations:
| Instance configuration type | Peak reads (QPS per region) | Peak writes (QPS total) | Peak writes using throughput optimized writes (QPS total) | ||
|---|---|---|---|---|---|
| Regional | 22,500 | or | 3,500 | 22,500 | |
| Dual-region and multi-region | 15,000 | or | 2,700 | 15,000 |
For information about the performance throughput of these instance configurations, see Performance for typical workloads.
Read guidance is given per region (because reads can be served from any read-write or read-only region), while write guidance is for the entire configuration. Read guidance assumes you're reading single rows of 1KB. Write guidance assumes that you're writing single rows at 1KB of data per row.
Peak write performance using throughput optimized writes is achieved using a batching delay of 100ms.
In general, both the read and write throughputs of a Spanner instance scale linearly as you add more compute capacity (nodes or processing units) to the instance. For example, if a single-region Spanner instance with 2 nodes can provide up to 45,000 reads per second, then a single-region Spanner instance with 4 nodes can provide up to 90,000 reads per second.
If you aren't seeing the expected performance for your workload from Spanner, see troubleshooting performance regressions for information about common causes.
Increased storage
For all Spanner regional, dual-region, and multi-region instance configurations, each node (1,000 processing units) of compute capacity in the instance has an increased storage capacity of 10 TB.
Performance for typical workloads
All Spanner instance configurations have improved performance and offer increased throughput.
Performance for regional configurations
Each 1,000 processing units (1 node) of compute capacity can provide the following peak performance (at 100% CPU) in a regional instance configuration:
| Peak reads (QPS per region) | Peak writes (QPS total) | Peak writes using throughput optimized writes (QPS total) | ||
|---|---|---|---|---|
| 22,500 | or | 3,500 | 22,500 |
For regional instance configurations that allow optional read-only replicas, the optional read-only replica can support an additional 5,000 reads per second.
Performance for dual-region configurations
Each 1,000 processing units (1 node) of compute capacity can provide the following peak performance (at 100% CPU) in a dual-region instance configuration. Use throughput optimized writes to increase write throughput beyond the numbers in the table.
| Base configuration name | Approximate peak reads (QPS per region) | Approximate peak writes (QPS total) |
|---|---|---|
dual-region-australia1 |
15,000 | 2,700 |
dual-region-germany1 |
15,000 | 2,700 |
dual-region-india1 |
15,000 | 2,700 |
dual-region-japan1 |
15,000 | 2,700 |
Read guidance is given per region (because reads can be served from anywhere), while write guidance is for the entire configuration. Read and write guidance assume that you're reading and writing single rows at 1 KB of data per row.
Performance for multi-region configurations
Each Spanner multi-region instance configuration has slightly different performance characteristics based on the replication topology. Use throughput optimized writes to increase write throughput beyond the numbers in the table.
Each 1,000 processing units (1 node) of compute capacity can provide the following peak performance (at 100% CPU):
| Base configuration name | Approximate peak reads (QPS per region) | Approximate peak writes (QPS total) |
|---|---|---|
asia1 |
15,000 | 2,700 |
asia2 |
15,000 | 2,700 |
eur3 |
15,000 | 2,700 |
eur5 |
15,000 | 2,700 |
eur6 |
15,000 7,500 for each optional read-only replica |
2,700 |
nam3 |
15,000 7,500 for each optional read-only replica |
2,700 |
nam6 |
15,000 in us-central1 and us-east1 7,500 in us-west1 and us-west2 [1] |
2,700 |
nam7 |
15,000 7,500 for each optional read-only replica |
2,700 |
nam8 |
15,000 | 2,700 |
nam9 |
15,000 | 2,700 |
nam10 |
15,000 | 2,700 |
nam11 |
15,000 7,500 for each optional read-only replica |
2,700 |
nam12 |
15,000 | 2,700 |
nam13 |
15,000 | 2,700 |
nam14 |
15,000 | 2,700 |
nam15 |
15,000 | 2,700 |
nam16 |
15,000 | 2,700 |
nam-eur-asia1 |
15,000 | 1,500 |
nam-eur-asia3 |
15,000 | 1,500 |
- [1]:
us-west1andus-west2provide only half of the QPS performance because they contain one replica per region instead of two.
Read guidance is given per region (because reads can be served from anywhere), while write guidance is for the entire configuration. Read and write guidance assume that you're reading and writing single rows at 1 KB of data per row.
Run your typical workloads against Spanner
Always run your own typical workloads against a Spanner instance
when doing capacity planning, so you can figure out the best resource allocation
for your applications. Google's PerfKit Benchmarker uses
YCSB to benchmark cloud services. You can
follow the PerfKitBenchmarker tutorial for Spanner
to create tests for your own workloads. When doing so, you should tune the
parameters in the benchmarking configuration yaml files to make
sure that the generated benchmark reflects the following characteristics in your
production environment:
- Total size of your database
- Schema (For example: row key size, number of columns, row data sizes)
- Data access pattern (row key distribution)
- Mixture of reads versus writes
- Type and complexity of queries
Reproduce benchmark numbers
To reproduce the benchmark numbers, follow the
Benchmarking Spanner with PerfKit Benchmarker tutorial
using the corresponding yaml files in the throughput_benchmark
folder.
To benchmark instances in an instance configuration that has undergone performance improvements, ensure that your tests are running in one of these improved instance configurations.
Zonal and regional failure protection
When running your workloads in production, it is important to provision enough compute capacity to continue to serve your traffic in the event of the loss of an entire zone (for regional instances) or an entire region (for dual-region and multi-region instances). For more information about the recommended maximum CPU, see alerts for high CPU utilization.
What's next
- Learn how to design a Spanner schema.
- Find out how to monitor Spanner performance.
- Learn how to troubleshoot issues with Key Visualizer.
- Learn about Spanner pricing.