This topic describes the capacity thresholds and queries per second (QPS) performance data of cloud-native gateways based on gateway specifications. This can help you select proper gateway specifications based on your business requirements. Gateway processing is a core workflow. Therefore, we recommend that you use security thresholds and warning thresholds together for your selection. In most cases, if a security threshold or warning threshold meet the selection requirements, the other threshold of the same specifications can also meet the selection requirements. In special business scenarios, for the same specifications, if the capacity threshold is met and QPS performance is not met or if QPS performance is met and the capacity threshold is not met, select higher specifications. This ensures that both of the two parameters can meet business requirements.
Capacity thresholds
The following table describes the capacity thresholds for a single gateway node based on gateway specifications. If the values of the gateway capacity metrics are less than the warning thresholds, your service-level agreement (SLA) requirements can be completely met. For your key business, we recommend that you keep the values of the gateway capacity metrics to be less than the security thresholds. This ensures the stability of your system.
Security thresholds: Security thresholds can ensure that a gateway maintains high throughput and low latency even when traffic is doubled.
Warning thresholds: If the usage reaches the warning thresholds, the latency of the gateway may increase, and traffic surges may result in stability risks.
Overload thresholds: If the current usage of a gateway exceeds the overload threshold, the gateway rejects new connections to improve its stability.
If a single node is deployed for your gateway, your SLA objective may not be met. You must make sure that at least two nodes are deployed for your gateway. If two or more nodes are deployed for your gateway, the security thresholds and warning thresholds are calculated based on the specifications of each node.
Gateway specifications | 2 cores and 4 GiB of memory | 4 cores and 8 GiB of memory | 8 cores and 16 GiB of memory | 16 cores and 32 GiB of memory | |
Number of nodes | 1 | 1 | 1 | 1 | |
Number of client connections | Security threshold | 12,000 | 24,000 | 48,000 | 96,000 |
Warning threshold | 24,000 | 48,000 | 96,000 | 192,000 | |
Overload threshold | 40000 | 80000 | 160000 | 320000 | |
Number of new HTTPS connections established per second | Security threshold | 400 | 800 | 1,600 | 3,200 |
Warning threshold | 800 | 1,600 | 3,200 | 6,400 | |
Overload threshold | - | - | - | - | |
Network bandwidth (Gbit/s) | Security threshold | 1 | 2 | 4 | 8 |
Warning threshold | 1 | 2 | 4 | 8 | |
Overload threshold | - | - | - | - | |
CPU utilization | Security threshold | 30% | 30% | 30% | 30% |
Warning threshold | 60% | 60% | 60% | 60% | |
Overload threshold | 90% | 90% | 90% | 90% | |
Memory usage | Security threshold | 75% | 75% | 75% | 75% |
Warning threshold | 75% | 75% | 75% | 75% | |
Overload threshold | 90% | 90% | 90% | 90% | |
QPS performance data
The QPS throughput of a gateway is affected by various factors, such as the response size, HTTPS support, and GZIP compression. The following table lists the QPS pessimistic values in worst cases when the CPU utilization of a gateway is less than 30%.
New HTTPS connections consume large amounts of CPU resources. In business scenarios in which a large number of HTTPS connections need to be established at the same time, you can estimate the gateway capacity based on the data of HTTPS short-lived connections, as shown in the following table.
Gateway specifications | 2 cores and 4 GiB of memory | 4 cores and 8 GiB of memory | 8 cores and 16 GiB of memory | 16 cores and 32 GiB of memory | |||||||
Number of nodes | 3 | 5 | 3 | 5 | 3 | 5 | 3 | 5 | |||
Connection type | Response size (KBytes) | HTTPS support | GZIP compression | QPS values when the CPU utilization is less than 30% | |||||||
Number of short-lived connections | 1 | No | No | 5,200 | 8,700 | 10,500 | 17,500 | 21,000 | 35,000 | 42,000 | 70,000 |
Yes | No | 1,600 | 2,700 | 3,200 | 5,500 | 6,500 | 11,000 | 13,000 | 22,000 | ||
Number of persistent connections | 1 | No | No | 6,500 | 10,800 | 13,000 | 21,700 | 26,000 | 43,500 | 52,000 | 87,000 |
Yes | No | 6,000 | 10,000 | 12,000 | 20,000 | 24,000 | 40,000 | 48,000 | 80,000 | ||
Yes | Yes | 5,200 | 8,700 | 10,500 | 17,500 | 21,000 | 35,000 | 42,000 | 70,000 | ||
10 | No | No | 5,600 | 9,300 | 11,200 | 18,700 | 22,500 | 37,500 | 45,000 | 75,000 | |
Yes | No | 5,300 | 9,000 | 10,700 | 18,000 | 21,500 | 36,000 | 43,000 | 72,000 | ||
Yes | Yes | 3,100 | 5,200 | 6,200 | 10,500 | 12,500 | 21,000 | 25,000 | 42,000 | ||