Yandex Data Streams pricing policy

To estimate your service costs, see the pricing in this section.

The prices for service products are also available in the price list.

All prices below do not include VAT.

Data Streams supports two pricing models:

• Based on allocated resources: You pay a fixed hourly rate for the established throughput limit and message retention period, and additionally for the number of units of actually written data.
• On-demand: You pay for the performed read/write operations, the amount of read or written data, and the actual storage used for messages that are still within their retention period.

The pricing model is set individually for each data stream. By default, new Data Streams data streams use the provisioned capacity pricing model.

Size of unit of written data is 40 KB.

Provisioned capacity pricing modelProvisioned capacity pricing model

In provisioned capacity mode, billing is based on allocated resources for data stream processing and the actually written data.

Pricing for provisioned resources is based on the allocated throughput limit per segment and storage configuration, i.e., size or retention period.

You get the first 744 hours of data storage and data stream maintenance with the following characteristics at no charge every month:

• Segment write speed: no more than 128 KBps.
• Message retention period: 1 hour.
• Number of data units written per month: no more than 2 million.

After you spend the free amount of services, you will be charged at the applicable rate. The unused balance of free services is reset at the end of the month.

Cost calculation examplesCost calculation examples

Let’s assume a single-shard stream receives two messages per second (50 KB each). The record retention period is 12 hours.

The cost of data processing per month (for a 31-day month) will be:

(50 / 40) = 2

Where:

• 50: Size of one message in KB.
• 40: Size of units of written data in KB.
• 2: Number of data units written per message.

While calculating the cost, the number of units of written data is rounded up to an integer. In this example, the number is rounded to 2.

2 × 2 = 4

Where:

• 2: Number of messages per second.
• 2: Number of data units written per message.
• 4: Number of data units written per second.

4 × 60 × 60 × 24 × 31 = 10,713,600

Where:

• 4: Number of data units written per second.
• 60 × 60 × 24 × 31: Number of seconds in the month.
• 10,713,600: Number of data units written per month.

$0.001200 × 24 × 31 = $0.892800

Where:

• $0.001200: Cost of allocated resources per hour when stored for 12 hours.
• 24: Number of data storage hours.
• 31: Number of days in a month.
• $0.892800: Cost of allocated resources per month with a maximum data write speed of 128 KB/s and a storage period of 24 hours, without VAT.

(10,713,600 - 2,000,000) / 1,000,000 × $0.010620 + $0.892800 = $0.985338

Where:

• 10,713,600: Number of data units written per month.
• 2,000,000: Number of free units per month.
• $0.010620: Cost of 1 million units of written data, without VAT.
• $0.892800: Cost of allocated resources per month with a maximum data write speed of 128 KB/s and a storage period of 24 hours, without VAT.

Total: $0.985338, without VAT.

If the data storage time is 1 hour, the time of resource usage by multiple streams is summed up. For example, if two streams are running 24/7 during a month, the time of resource usage will be 2 × 24 × 30 = 1.440 hours. The first 744 hours are free of charge. They will be used up by the two streams in half a month. After that, you will start to be charged for the service.

With extended storage time, data is stored for up to 7 days.

With the storage limit provided, the minimum storage capacity is 50 GB per segment.

On-demand pricingOn-demand pricing

With on-demand pricing:

• Data streams stored in YDB serverless databases are charged for based on the pricing policy YDB for the serverless mode.

• Data streams stored in YDB dedicated databases are not charged for separately. Instead, you only pay for a dedicated database (see the pricing policy for dedicated databases for details).