Cloud Computing 17 min read

How to Choose the Right Alibaba Cloud ECS Instance and Validate Capacity with PTS

This guide explains how enterprises can select appropriate Alibaba Cloud ECS instance types using spec families or API queries, match them to workloads such as web services, big data, or AI, and verify capacity and performance through systematic PTS load testing and capacity planning.

Alibaba Cloud Native
Alibaba Cloud Native
Alibaba Cloud Native
How to Choose the Right Alibaba Cloud ECS Instance and Validate Capacity with PTS

01. Selecting by Instance Specification Parameters

Before launching an ECS instance, consider performance, price, and workload to choose the right configuration. Alibaba Cloud offers multiple instance families. You can find suitable specs via two methods:

Instance specification families – consult the documentation for detailed product information. DescribeInstanceTypes API – call the ECS API to retrieve the latest performance parameters.

Instance names follow the pattern ecs.<family> for families and ecs.<family>.<nx>large for specific sizes, where nx indicates the number of vCPU cores (e.g., xlarge = 4 cores, 2xlarge = 8 cores).

Families include x86 and ARM compute families, heterogeneous compute families, elastic bare‑metal servers, and SCC (super‑computer) families, each using a mixed alphanumeric naming scheme.

02. Selecting by Self‑Built Services and Applications

Enterprises often build custom services to meet business needs. The guide maps common services to recommended ECS families:

General web services – balanced CPU:memory ratio of 1:2, high‑performance cloud disks (SSD or ESSD).

Big‑data workloads (Hadoop, Spark, Kafka) – use ecs.g6e families for compute nodes and d series for storage‑intensive data nodes (e.g., ecs.d2s.5xlarge for MapReduce).

Databases and caches – prioritize CPU:memory ratios >1:4 and high IOPS storage; lightweight databases use cost‑effective instances, heavy‑load databases benefit from i‑series with local NVMe SSD.

Deep learning and image processing – require GPU accelerators; recommended GPU:CPU ratios range from 1:4 to 1:12 depending on training or inference workloads.

03. Selecting by Application Scenarios

Beyond individual services, many business scenarios combine multiple components. For general workloads, choose instances with balanced CPU and memory (1:2) and high‑speed cloud disks; for video streaming or high‑network demand, select higher‑spec instances within the same family to increase PPS.

For database, cache, or search scenarios, aim for CPU:memory ratios >1:4 and prioritize storage I/O performance.

04. Specification Verification and Capacity Planning

After selecting an ECS instance, monitor resource utilization. When usage is below saturation, TPS grows linearly with concurrency; once resources saturate, CPU spikes and TPS plateaus, potentially leading to capacity collapse.

Different teams focus on different metrics:

Business team – can the capacity sustain peak activity?

R&D team – are manual or automatic scaling, deployments, and rollbacks unrestricted by resources?

Operations team – resource utilization, departmental quotas, and cost.

Traditional manual sizing often over‑provisions resources. Use Alibaba Cloud Performance Testing Service (PTS) to validate instance choices, identify safe, limit, and breaking capacity points, and implement throttling or degradation strategies.

Using PTS for Load Testing

Activate free‑trial PTS and ECS if eligible; the trial provides 5,000 VUM free for the first month.

Configure a one‑click ROS template to create a security group, deploy a LAMP stack, and install PbootCMS.

Follow the ROS wizard: select region, specify the ECS instance, set MySQL root password, and create the stack.

After stack creation, retrieve the public IP and access http://<ECS_IP>/admin.php to verify deployment.

In the PTS console, create a new scene, set the target URL (e.g., http://<ECS_IP>:80), and configure load parameters (auto‑increment, max concurrency 50, step 10%, duration 5 min).

Start the test, monitor real‑time metrics (success rate, response time, TPS), and after completion view the detailed report.

Follow the free‑trial limits to avoid unexpected charges.

Further Capacity Analysis

Post‑test, analyze bottlenecks across security, load balancers, backend services, middleware, databases, OS, and hardware. Upcoming articles will detail tuning steps.

Relevant links:

Instance specification families: https://help.aliyun.com/document_detail/25378.htm#concept-sx4-lxv-tdb

DescribeInstanceTypes API: https://help.aliyun.com/document_detail/25620.htm#doc-api-Ecs-DescribeInstanceTypes

Performance Testing Service: https://www.aliyun.com/product/pts

PTS pricing rules: https://help.aliyun.com/document_detail/433167.html?spm=a2c4g.29269.0.0.67fa7f32p02i2O

ECS console: https://account.aliyun.com/login/login.htm?oauth_callback=https%3A%2F%2Fecs.console.aliyun.com%2F

One‑click ROS template: https://account.aliyun.com/login/login.htm?oauth_callback=https%3A%2F%2Fros.console.aliyun.com%2Fregion%2Fstacks%2Fcreate%3Fspm%3Da2c4g.611918.0.0.3aec628amGQK9n%26templateUrl%3Dhttps%3A%2F%2Fstatic-aliyun-doc.oss-cn-hangzhou.aliyuncs.com%2Ffile-manage-files%2Fzh-CN%2F20230320%2Fonyv%2F%25E5%25BF%25AB%25E9%2580%259F%25E6%2590%25AD%25E5%25BB%25BA%25E7%25BD%2591%25E7%25AB%2599.yml&lang=zh

PTS console: https://account.aliyun.com/login/login.htm?oauth_callback=https%3A%2F%2Fpts.console.aliyun.com%2F&lang=zh

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

cloud computingPerformance Testingcapacity planningECSInstance Selection
Alibaba Cloud Native
Written by

Alibaba Cloud Native

We publish cloud-native tech news, curate in-depth content, host regular events and live streams, and share Alibaba product and user case studies. Join us to explore and share the cloud-native insights you need.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.