MSE Nacos Enterprise Edition supports a 99.99% Service-Level Agreement (SLA) and improves service push performance by 300% compared to a self-managed open source Nacos instance. It offers higher quota capabilities using dedicated underlying core resources. This topic describes the capacity thresholds and provides queries per second (QPS) performance references for different specifications of Nacos Enterprise Edition.
Version overview
Capacity thresholds
Connection assessment: This is the number of connections established between Nacos clients and the Nacos server. In a typical scenario, one pod establishes one connection. However, if you use both the service registry and the configuration center, two connections are required.
Level assessment: For core business applications, keep capacity metrics below the safe level. This practice ensures high stability even during sudden traffic bursts. If the QPS or Transaction Per Second (TPS) exceeds the alert level, traffic bursts may threaten stability. You should also keep capacity metrics below the safe level when you restart or upgrade a cluster.
Assessment standard: To ensure stability, the following capacity assessments are based on client version 1.x. For better performance, we recommend that you upgrade your client to version 2.x.
The following table shows the capacity metrics for a 3-node cluster.
To estimate the capacity of clusters with a different node count, you can proportionally scale the data from a 3-node cluster of the same specification.
Specification | Number of nodes | Connections | Queries per second (QPS) | Transactions per second (TPS) | |||
Safe level | Alert level | Safe level | Alert level | Safe level | Alert level | ||
Small.Platinum.x2 | 3 | 1200 | 1800 | 1200 | 1800 | 600 | 900 |
Medium.Platinum.x1 | 3 | 2400 | 3600 | 2400 | 3600 | 1200 | 1800 |
Medium.Platinum.x2 | 3 | 4800 | 7200 | 4800 | 7200 | 2400 | 3600 |
Large.Platinum.x1 | 3 | 9600 | 14400 | 9600 | 14400 | 4800 | 7200 |
Antifragility throttling
To ensure stability, MSE Nacos Enterprise Edition enables antifragility mode by default. When the resource capacity or API traffic reaches a threshold, the system automatically applies throttling and capacity controls. The following table shows the throttling rules for a single node.
To determine the throttling rules for a multi-node cluster, you can proportionally scale the data from a single node.
Specification | Connections | Publications of the same configuration (times/minute) | Configuration publications (TPS) | Configuration queries (QPS) | Service pushes (TPS) |
Small.Platinum.x2 | 800 | 20 | 100 | 200 | 800 |
Medium.Platinum.x1 | 1600 | 20 | 100 | 400 | 1600 |
Medium.Platinum.x2 | 3200 | 20 | 100 | 800 | 3200 |
Large.Platinum.x1 | 6400 | 20 | 100 | 1600 | 6400 |