This document describes the bandwidth billing effective time, free tier, billing prices, metering rules and settlement cycle, billing formulas and examples, bandwidth lifecycle and account overdue payment handling, as well as the impact and operational recommendations of disabling additional bandwidth for Tencent Cloud Distributed Cache (Redis-compatible).
Billing Overview
Tencent Cloud Distributed Cache provides standard bandwidth (free bandwidth) based on the instance architecture and memory specification. No fees are charged when bandwidth usage does not exceed the standard bandwidth. The excess portion is metered on an hourly basis and billed at the unit price of the region where the instance resides.
Effective Time
Bandwidth billing takes effect in batches by region, and the effective time for each region is based on Beijing time:
Starting from 2026-12-01 00:00:00, additional bandwidth that exceeds the standard bandwidth in regions outside the Chinese mainland will be officially billed.
Starting from 2026-12-08 00:00:00, additional bandwidth that exceeds the standard bandwidth in the Chinese mainland will be officially billed.
Before the effective time for each region, additional bandwidth that exceeds the standard bandwidth will not be billed.
Bandwidth Free Tier
Tencent Cloud Distributed Cache provides standard bandwidth of different amounts based on the instance architecture and memory specification. The following table lists the standard bandwidth amount for each specification, in Mb/s (megabits per second).
|
Standard architecture | Less than 1 | 80 | 10 |
Standard architecture | 1 | 128 | 16 |
Standard architecture | (1, 8] | 192 | 24 |
Standard architecture | (8, 16] | 256 | 32 |
Standard architecture | (16, 32] | 384 | 48 |
Standard architecture | (32, 64] | 512 | 64 |
Cluster Architecture | — | 768 | 96 |
Note:
The interval notation (1, 8] means greater than 1 GB and less than or equal to 8 GB, and the same applies to other intervals.
Standard bandwidth is the complimentary quota for each node (shard). The total standard bandwidth of an instance = standard bandwidth × number of shards, and the number of shards in the standard architecture is 1. For example, the total standard bandwidth of a cluster architecture instance = 768 Mb/s × number of shards, which is 768 ÷ 8 × number of shards = 96 MB/s × number of shards when converted to MB/s. For a cluster architecture instance with 3 shards, the total standard bandwidth = 768 Mb/s × 3 = 2304 Mb/s, which is 2304 ÷ 8 = 288 MB/s when converted.
The general formula for calculating the total standard bandwidth of an instance is: standard bandwidth × number of shards × Max(read-only replica quantity, 1). When read-only replicas are not enabled, the read-only replica quantity is counted as 0 and the coefficient is 1, which is the formula above. After read-only replicas are enabled, each read-only replica has the same bandwidth quota as the master node. For the calculation of the total instance bandwidth (including additional bandwidth), see Bandwidth Adjustment. Bandwidth Pricing
Bandwidth usage that exceeds the free tier is billed at different unit prices by region. The following table compares the bandwidth prices for the two types of regions.
|
Chinese mainland region | USD 0.003044/(MB/s)/hour |
Non-Chinese mainland region | USD 0.003696/(MB/s)/hour |
Note:
The billable bandwidth rate is rounded up to the nearest 1 MB/s, and any amount less than 1 MB/s is calculated as 1 MB/s. If the billable duration is less than 1 hour, it is calculated as 1 hour.
Metering Rules and Billing Cycles
Additional bandwidth is billed based on actual usage, and charges are incurred only for the portion that exceeds the standard bandwidth. No bandwidth charges are incurred for hours that do not exceed the standard bandwidth. The rules for measuring, determining, and settling bandwidth are as follows:
|
Measurement unit | Bandwidth is billed by rate. The billable usage is the cumulative value of "rate × duration", measured in (MB/s)·hour. |
Statistical Cycle | Billing is based on hourly cycles. Each hour generates one billable usage record (the billable bandwidth rate of that hour × 1 hour), which is aggregated and settled at the unit price of the region where the instance resides. The additional bandwidth usage for a calendar day is the sum of 24 hourly billable usage records, and for a 31-day calendar month, it is the sum of 744 hourly billable usage records. |
Bandwidth value determination | Based on the second-level data collected by system monitoring, take the maximum inbound and outbound traffic rates within the hour, and then use the larger of the two as the bandwidth usage rate (MB/s) for that hour. The bandwidth usage rate is determined solely by the peak value within the hour, regardless of the average rate, cumulative transfer volume, or peak duration. |
Inbound and outbound directions | Both inbound traffic (such as write requests entering the instance) and outbound traffic (such as read responses leaving the instance) are included in the calculation. The peak values of the two directions are taken separately, and the larger one is used. The two directions are not added together, nor are they billed separately. |
Collection granularity | Instance monitoring data supports granularities of 5 seconds, 1 minute, 5 minutes, 1 hour, and 1 day. Both the console and API support queries at these granularities, and billing values are consistent with system monitoring data. The minimum granularity actually supported by an instance depends on the monitoring granularity version of the instance. For details, see View Monitoring Data (Redis/Valkey edition). |
Settlement and deduction | Additional bandwidth fees are independent of the instance purchase cost. For yearly/monthly subscription instances, the instance fee is settled by prepaid billing cycle. Additional bandwidth is metered and billed on an hourly basis. After each hour ends, the system settles the billable usage of that hour at the regional unit price and deducts it from the account balance. The hourly accumulation is the total additional bandwidth fee within the billing cycle. Bill details can be viewed in the Billing Center. |
Billing Formulas and Examples
Additional bandwidth charges are calculated only for the portion that exceeds the free tier, and are accumulated hour by hour on an hourly basis. In the formula, bandwidth is measured by rate in MB/s, and the billing duration is measured in hours:
Free bandwidth rate = total standard bandwidth of the instance ÷ 8 (converting Mb/s to MB/s)
Hourly billable bandwidth rate = ⌈hourly bandwidth usage rate - free bandwidth rate⌉ (rounded up, and if the result is less than or equal to 0, it is counted as 0)
Hourly billable usage (unit: MB/s·hour) = hourly billable bandwidth rate × 1 hour
Additional bandwidth cost = Σ hourly billable usage × bandwidth unit price
In the formula, ⌈ ⌉ indicates rounding up to the nearest 1 MB/s: if the excess is greater than 0 but less than 1 MB/s, it is calculated as 1 MB/s; if the excess contains a decimal, it is rounded up to the next integer, for example, an excess of 16.3 MB/s is calculated as 17 MB/s. If the bandwidth in an hour does not exceed the free tier, no billable usage is generated for that hour.
Take a standard architecture instance in the Guangzhou region as an example. Assume that read-only replicas are not enabled for this instance, and it is configured with 1 master node and 1 standby node, with 1 GB per node. In a given hour, the peak inbound traffic rate is 32 MB/s and the peak outbound traffic rate is 27 MB/s. The larger of the two values, 32 MB/s, is used as the bandwidth usage rate for that hour. The billing process is as follows:
Free bandwidth rate: For a 1 GB standard architecture, the standard bandwidth of a single node is 128 Mb/s. With 1 shard and read-only replicas not enabled, the coefficient is 1. The total standard bandwidth of the instance = 128 Mb/s × 1 × 1 = 128 Mb/s, which is converted to 128 ÷ 8 = 16 MB/s. Replica nodes are not separately counted toward the free tier.
Billable bandwidth rate: 32 MB/s - 16 MB/s = 16 MB/s.
Billable usage for the hour: 16 (MB/s) × 1 hour = 16 (MB/s)·hour.
Additional bandwidth cost for the hour: 16 (MB/s)·hour × USD 0.003044/(MB/s)/hour = USD 0.048704.
This example assumes that read-only replicas are not enabled for the instance. If read-only replicas are enabled, recalculate the total standard bandwidth of the instance by using the formula in the free bandwidth quota section of this document, and then substitute it into the preceding steps.
The preceding example shows the cost for a single full hour. If the instance generates usage continuously over multiple hours, calculate the billable usage for each hour and add up the results. Multiply the accumulated result by the regional unit price to obtain the total additional bandwidth cost for the billing cycle.
Bandwidth Lifecycle
During normal operation of an instance, the bandwidth of instances in both the yearly/monthly subscription and pay-as-you-go billing modes can be used normally, and the portion exceeding the standard bandwidth is billed according to the rules in this document. After an instance expires or has overdue payments, whether the bandwidth remains available and whether billing continues depend on the billing mode and the current stage. The specific differences are as follows:
|
Grace period after expiration or overdue payments | Within 7 days after the instance expires, the bandwidth can be used normally, and the portion exceeding the standard bandwidth is still billed. | Within 24 hours after the instance has overdue payments, the bandwidth can be used normally, and the portion exceeding the standard bandwidth is still billed. |
Isolation to recycle bin | Starting from the 8th day after the instance expires, it is isolated and moved to the recycle bin, and no longer incurs bandwidth usage or fees. | After the instance has overdue payments for 24 hours, it is isolated and moved to the recycle bin, and no longer incurs bandwidth usage or fees. |
Disabling Additional Bandwidth
Billing Impact After Disabling
To avoid additional bandwidth charges, you can disable additional bandwidth for the target instance in the console. After the disabling takes effect, the instance uses only the standard bandwidth (free bandwidth) corresponding to its current specification, and no longer uses additional bandwidth or incurs additional bandwidth charges. Note that disabling additional bandwidth does not waive charges incurred before the disabling takes effect, and such charges are still settled according to the actual billing rules. Disabling additional bandwidth affects only the usage and billing of additional bandwidth, and does not affect the instance itself or the billing of other billable items.
Business Rate Limiting Risks
Warning:
Disabling additional bandwidth reduces the maximum available bandwidth of the instance. When business bandwidth demand exceeds the standard bandwidth quota, the business may be throttled, which can lead to increased request latency, reduced throughput, or request timeouts. Fully evaluate the business impact and proceed with caution. In particular, avoid disabling additional bandwidth directly during peak business hours without prior evaluation.
Operation Suggestion
The key considerations before, during, and after disabling additional bandwidth differ. We recommend that you perform the operations in the following order:
1. Evaluate before disabling: Check recent bandwidth monitoring data and peak business usage to determine whether the standard bandwidth can meet daily access, burst traffic, and future business growth requirements. For cluster architecture instances, also monitor the bandwidth usage of each shard to avoid making judgments based solely on the total instance usage.
2. Operate during off-peak hours: After confirming that the risks are acceptable, disable additional bandwidth during off-peak hours and prepare for restoration.
3. Observe after disabling: Continuously monitor bandwidth utilization, request latency, timeouts, and errors. If throttling or business performance degradation occurs, promptly evaluate re-enabling additional bandwidth. After additional bandwidth is re-enabled, usage exceeding the standard bandwidth is billed according to the billing rules in this document.
Cost and performance must be balanced: disabling additional bandwidth can avoid future additional bandwidth charges, but it also reduces the maximum available bandwidth of the instance. If the standard bandwidth cannot cover peak business demand, we recommend that you keep additional bandwidth enabled, or optimize business traffic before evaluating whether to disable it.