Serverless instances are billed on a pay-as-you-go basis so you pay only for the resources you actually consume. This topic describes the billing items, unit prices, free quotas, and worked examples for Serverless instances, and helps you estimate cost for your workload.
Applicable scope
This topic applies to ApsaraMQ for RocketMQ Serverless instances. To confirm your instance type, open the ApsaraMQ for RocketMQ console and go to Instances > Instance Details > Basic Information. If the Primary Series is Standard or Professional, this topic does not apply — see Billing of Standard and Professional edition instances instead.
Serverless instances use one of two billing types, and each billing item applies to only one or both of them:
Billing item | Cumulative pay-as-you-go | Reserved + elastic |
Message sending and receiving fees | Yes | No |
Instance fees | Yes | Yes |
Reserved capacity fees | No | Yes |
Elastic TPS fees | No | Yes |
Online client count fees | Yes | Yes |
Resource fees | Yes | Yes |
Storage fees | Yes | Yes |
Public network fees | Yes (only when internet access is enabled) | Yes (only when internet access is enabled) |
For guidance on the differences between the legacy Standard and Professional editions and the new cumulative pay-as-you-go model, how to identify your instance type, and how to migrate to the shared edition, see Migration guide to the shared edition.
Usage limits
Internet access adaptation — When you access a Serverless instance over the internet, additional SDK code adaptation is required. For details, see Internet access for Serverless instances.
Billing items
Serverless instances start incurring charges at the time of purchase, and bills are generated hourly under the pay-as-you-go model. The following diagram shows the billing items covered in this topic. For the mapping of each item to a billing type, see the table under Applicable scope.
Message sending and receiving fees
Message sending and receiving fees apply to Serverless instances that use cumulative pay-as-you-go billing.
Billing formula
Message sending and receiving fees = Cumulative number of messages sent and received (times) × Unit price per message (USD/time)
Calculation rules
The number of messages sent and received equals the number of messages received plus the number of messages delivered.
Messages received — The number of messages the server accepts from producers. If a topic receives 5 messages, the count is 5.
Messages delivered — The number of times the server delivers messages to consumers. Each successful delivery counts as 1, regardless of whether the consumer processes the message or returns an ACK response.
Payload size adjustment — The count is multiplied by a payload factor based on message size, using 4 KB as the base unit and rounding up. For example, delivering an 18 KB message counts as ⌈18/4⌉ = 5 deliveries. ⌈ ⌉ denotes rounding up.
Tiered unit prices
Unit prices for message sending and receiving vary by region. Values below are per million messages.
Public cloud regions (USD per million messages)
Region | Billing tier | Monthly cumulative messages (× 10⁸) | Send-and-consume unit price (normal messages) | Consume unit price (scheduled, delayed, or transaction messages) |
China (Hangzhou), China (Shanghai), China (Shenzhen), China (Qingdao), China (Beijing), China (Zhangjiakou), China (Hohhot), China (Chengdu), China (Ulanqab), China (Heyuan), China (Guangzhou), Singapore, Germany (Frankfurt), UK (London), US (Silicon Valley), US (Virginia) | Tier 1 | (0, 10] | 0.305 | 1.525 |
Tier 2 | (10, 50] | 0.183 | 0.915 | |
Tier 3 | (50, 200] | 0.153 | 0.765 | |
Tier 4 | >200 | 0.122 | 0.61 | |
UAE (Dubai) | Tier 1 | (0, 10] | 0.61 | 3.05 |
Tier 2 | (10, 50] | 0.366 | 1.83 | |
Tier 3 | (50, 200] | 0.306 | 1.53 | |
Tier 4 | >200 | 0.244 | 1.22 | |
US (Atlanta) | Tier 1 | (0, 10] | 0.412 | 2.059 |
Tier 2 | (10, 50] | 0.247 | 1.235 | |
Tier 3 | (50, 200] | 0.207 | 1.033 | |
Tier 4 | >200 | 0.165 | 0.824 | |
Malaysia (Johor) | Tier 1 | (0, 10] | 0.411 | 2.055 |
Tier 2 | (10, 50] | 0.247 | 1.235 | |
Tier 3 | (50, 200] | 0.206 | 1.03 | |
Tier 4 | >200 | 0.165 | 0.825 |
Riyadh region (USD per million messages)
Region | Billing tier | Monthly cumulative messages (× 10⁸) | Send-and-consume unit price (normal messages) | Consume unit price (scheduled, delayed, or transaction messages) |
Tier 1 | (0, 10] | 0.366 | 1.83 | |
Tier 2 | (10, 50] | 0.2196 | 1.098 | |
Tier 3 | (50, 200] | 0.1836 | 0.918 | |
Tier 4 | >200 | 0.1464 | 0.732 |
Instance fees
Instance fees apply to all Serverless instances, regardless of billing type. The fee covers PrivateLink instances in two availability zones and includes free quotas for online clients and resources. For the specific free quota values, see Free quota for online clients and Free quota for resources in this topic.
Unit prices
Region | Unit price (USD/instance/hour) |
All regions | 0.0452 |
Billing example
For a full month of 30 days: 0.0452 × 24 × 30 = 32.544 USD.
Reserved capacity fees
Reserved capacity fees apply to Serverless instances that use reserved + elastic billing. The fee is calculated from the reserved capacity you specify and the duration for which the capacity is held.
Billing formula
Hourly reserved capacity fee = Reserved capacity (TPS) × Reserved capacity unit price (USD/TPS/hour)
Unit prices
Public cloud regions
Region | Billing tier | Shared: reserved capacity (TPS) | Shared: unit price (USD/TPS/hour) | Dedicated: reserved capacity (TPS) | Dedicated: unit price (USD/TPS/hour) |
China (Hangzhou), China (Shanghai), China (Shenzhen), China (Qingdao), China (Beijing), China (Zhangjiakou), China (Hohhot), China (Chengdu), China (Ulanqab), China (Heyuan), China (Guangzhou), Singapore, Germany (Frankfurt), UK (London), US (Silicon Valley), US (Virginia) | Tier 1 | (0, 2000] | 0.000135 | (0, 5000] | 0.000173 |
Tier 2 | (2000, 8000] | 0.000113 | (5000, 20000] | 0.000128 | |
Tier 3 | (8000, 50000] | 0.00006 | (20000, 100000] | 0.000098 | |
Tier 4 | >50000 | 0.000045 | >100000 | 0.000075 | |
UAE (Dubai) | Tier 1 | (0, 2000] | 0.00027 | (0, 5000] | 0.000346 |
Tier 2 | (2000, 8000] | 0.000226 | (5000, 20000] | 0.000256 | |
Tier 3 | (8000, 50000] | 0.00012 | (20000, 100000] | 0.000196 | |
Tier 4 | >50000 | 0.00009 | >100000 | 0.00015 | |
US (Atlanta) | Tier 1 | (0, 2000] | 0.000182 | (0, 5000] | 0.000234 |
Tier 2 | (2000, 8000] | 0.000153 | (5000, 20000] | 0.000173 | |
Tier 3 | (8000, 50000] | 0.000081 | (20000, 100000] | 0.000132 | |
Tier 4 | >50000 | 0.000061 | >100000 | 0.000101 | |
Malaysia (Johor) | Tier 1 | (0, 2000] | 0.000182 | (0, 5000] | 0.000234 |
Tier 2 | (2000, 8000] | 0.000152 | (5000, 20000] | 0.000173 | |
Tier 3 | (8000, 50000] | 0.000081 | (20000, 100000] | 0.000132 | |
Tier 4 | >50000 | 0.00006 | >100000 | 0.0001 |
Riyadh region
Region | Billing tier | Shared: reserved capacity (TPS) | Shared: unit price (USD/TPS/hour) | Dedicated: reserved capacity (TPS) | Dedicated: unit price (USD/TPS/hour) |
Tier 1 | (0, 2000] | 0.000162 | (0, 5000] | 0.0002076 | |
Tier 2 | (2000, 8000] | 0.0001356 | (5000, 20000] | 0.0001536 | |
Tier 3 | (8000, 50000] | 0.000072 | (20000, 100000] | 0.0001176 | |
Tier 4 | >50000 | 0.000054 | >100000 | 0.00009 |
Billing example
Assume a Serverless instance in China (Hangzhou) uses the shared deployment architecture with reserved + elastic billing, and its reserved capacity is 4,000 TPS. The hourly reserved capacity fee is: 2000 × 0.000135 + (4000 - 2000) × 0.000113 = 0.496 USD. For a full month of 30 days, the total is: 0.496 × 24 × 30 = 357.12 USD.
Capacity selection reference
Size your reserved capacity based on actual concurrent traffic rather than theoretical peak. The following reserved capacity recommendations apply to common scenarios.
Small e-commerce systems — A reserved capacity of 500 to 2,000 TPS is typically sufficient.
E-commerce with about 100,000 concurrent users per hour and a QPS of 2,000 to 3,000 — A reserved capacity of 2,000 TPS covers daily traffic.
Traffic that exceeds reserved capacity — Excess traffic is automatically billed at the elastic TPS rate. See Elastic TPS fees in this topic. You do not need to reserve extra capacity for infrequent spikes; size reserved capacity to your daily peak to avoid paying for capacity you rarely use.
Elastic TPS fees
Elastic TPS fees apply to Serverless instances that use reserved + elastic billing, and are charged for traffic that exceeds the reserved capacity.
Billing formula
`
Hourly elastic TPS fee = ∑(1≤i≤60) [Ti × Elastic TPS] × Elastic TPS unit price (CNY/TPS/minute)`
`
Hourly elastic TPS fee = ∑(1≤i≤60) [Ti × Elastic TPS] × Elastic TPS unit price (USD/TPS/minute)`
Where T1, T2, ..., Ti, ..., T60 represent the 1st minute, 2nd minute, ..., i-th minute, ..., 60th minute of the hour, respectively.
∑ represents the total elastic TPS used within the hour.
Unit prices
Public cloud regions
Region | Shared unit price (USD/TPS/hour) | Dedicated unit price (USD/TPS/hour) |
China (Hangzhou), China (Shanghai), China (Shenzhen), China (Qingdao), China (Beijing), China (Zhangjiakou), China (Hohhot), China (Chengdu), China (Ulanqab), China (Heyuan), China (Guangzhou), Singapore, Germany (Frankfurt), UK (London), US (Silicon Valley), US (Virginia) | 0.000019 | 0.000025 |
UAE (Dubai) | 0.000038 | 0.00005 |
US (Atlanta) | 0.000026 | 0.000034 |
Malaysia (Johor) | 0.000026 | 0.000034 |
Riyadh region
Region | Shared unit price (USD/TPS/hour) | Dedicated unit price (USD/TPS/hour) |
0.0000228 | 0.00003 |
Billing example
Assume a Serverless instance in China (Hangzhou) uses the shared deployment architecture with reserved + elastic billing, and its reserved capacity is 5,000 TPS. During one hour, the instance runs as follows.
Time | Minute 1 | Minute 2 | Minute 3 | Minute 4 | ...... | Minute 59 | Minute 60 |
Actual TPS traffic peak of the instance (times/second) | 5200 | 5300 | 4000 | 4500 | All below 5000 | 4800 | 5100 |
Incremental TPS (times/second) | 200 | 300 | 0 | 0 | 0 | 0 | 100 |
The elastic TPS fee for this hour is: (1 × 200 + 1 × 300 + 1 × 100) × 0.000019 = 0.0114 USD.
Online client count fees
Online client count fees apply to all Serverless instances. When the number of online clients exceeds the free quota, the excess counts as paid connections and is billed hourly (any partial hour is counted as a full hour).
Client counting rules
One producer or one consumer object counts as one online client.
Billing formula
Online client count fee = Number of paid online clients (count) × Online client unit price (USD/client/hour)
Free quota for online clients
Deployment architecture | Capacity mode | TPS specification | Maximum online clients | Free online clients |
Shared | Cumulative pay-as-you-go | All | 5000 | 500 |
Shared | Reserved + elastic | [2000, 20000] | 5000 | 1000 |
Shared | Reserved + elastic | (20000, 50000] | 8000 | 2000 |
Shared | Reserved + elastic | (50000, 100000] | 10000 | 3000 |
Shared | Reserved + elastic | (100000, 200000] | 20000 | 5000 |
Dedicated | Reserved + elastic | 5000 | 3000 | 2000 |
Dedicated | Reserved + elastic | 10000 | 6000 | 4000 |
Dedicated | Reserved + elastic | 15000 | 8000 | 6000 |
Dedicated | Reserved + elastic | [20000, 50000] | 10000 | 8000 |
Dedicated | Reserved + elastic | (50000, 100000] | 20000 | 10000 |
Dedicated | Reserved + elastic | (100000, 200000] | 40000 | 20000 |
Dedicated | Reserved + elastic | (200000, 300000] | 80000 | 40000 |
Dedicated | Reserved + elastic | (300000, 500000] | 100000 | 50000 |
Dedicated | Reserved + elastic | (500000, 1000000] | 200000 | 100000 |
Unit prices
Region | Unit price (USD/client/hour) |
All regions | 0.00003 |
Resource fees
Resource fees apply to all Serverless instances. Resources refers to the total number of topics and consumer groups. When the resource count exceeds the free quota, the excess counts as paid resources and is billed hourly (any partial hour is counted as a full hour).
Billing formula
Resource fee = Number of paid resources (count) × Resource unit price (USD/resource/hour)
Free quota for resources
Deployment architecture | Capacity mode | TPS specification | Free resource quota (count) |
Shared | Cumulative pay-as-you-go | All | 100 |
Shared | Reserved + elastic | All | 200 |
Dedicated | Reserved + elastic | [0, 10000] | 300 |
Dedicated | Reserved + elastic | (10000, 100000] | 500 |
Dedicated | Reserved + elastic | (100000, 200000] | 1000 |
Dedicated | Reserved + elastic | (200000, 300000] | 1500 |
Dedicated | Reserved + elastic | (300000, 500000] | 2000 |
Dedicated | Reserved + elastic | (500000, 1000000] | 3000 |
Unit prices
Region | Billing tier | Resource count (exceeding free quota) | Unit price (USD/resource/hour) |
All regions | Tier 1 | [0, 500] | 0.00015 |
All regions | Tier 2 | (500, 1000] | 0.000092 |
All regions | Tier 3 | >1000 | 0.00006 |
Storage fees
Storage fees apply to all Serverless instances, and are calculated from the storage space occupied by messages on the server and the storage duration. Any storage duration less than 1 hour is counted as 1 hour.
Billing formula
Hourly storage fee = Storage space used (GB) × Storage unit price (USD/GB/hour)
Unit prices
Region | Storage unit price (USD/GB/hour) |
China (Hangzhou), China (Shanghai), China (Shenzhen), China (Qingdao), China (Beijing), China (Zhangjiakou), China (Hohhot), China (Chengdu), China (Ulanqab), China (Heyuan), China (Guangzhou), Singapore, Germany (Frankfurt), UK (London), US (Silicon Valley), US (Virginia), US (Atlanta), Malaysia (Johor) | 0.0002 |
UAE (Dubai) | 0.00036 |
Public network fees
Public network fees apply when you access Serverless instances over the internet. Fees are calculated from outbound public network traffic. If internet access is not enabled, no public network outbound traffic fees are charged.
Billing formula
Hourly public network fee = Cumulative outbound public network traffic (GB) × Traffic unit price (USD/GB)
Unit prices
Region | Outbound public network traffic unit price (USD/GB/hour) |
China (Hangzhou), China (Shanghai), China (Shenzhen), China (Qingdao), China (Beijing), China (Zhangjiakou), China (Hohhot), China (Chengdu), China (Ulanqab), China (Heyuan), China (Guangzhou), Singapore, Germany (Frankfurt), UK (London), US (Silicon Valley), US (Virginia), UAE (Dubai), US (Atlanta), Malaysia (Johor) | 0.15 |
0.19 |