With Amazon Timestream you only pay for what you use, there are no up-front costs, and there is no minimum fee. Amazon Timestream charges for usage based on the following dimensions:

  • Writes: The amount of data written from your applications (rounded to the nearest KB) into a table.
  • Queries: The amount of data scanned by Amazon Timestream’s serverless distributed query engine while computing query results (rounded to the nearest MB, with a 10 MB minimum).
  • Memory store: The amount of data stored in the memory store of each table.
  • Magnetic store: The amount of data stored in the magnetic store of each table.
  • SSD store: Coming soon.
 
Amazon Timestream also offers the following features to help you optimize costs (see  Metering and Cost Optimization for more details):

  • Multi-measure time-series records: Multi-measure time-series records help optimize the cost of data writes, storage, and queries by storing multiple measures per table row.
  • Data batching: Batching multiple records in a single write request helps reduce the cost of writes. See batching time-series events for details.
  • Automatic detection of repeated data: When applications batch multiple records while writing data into Amazon Timestream, the service automatically identifies repeatedly occurring dimensions in the data and charges only once for repeated dimensions. This further helps optimize the cost of writes.
  • Magnetic store writes: Writing late arrival data (data with a timestamp in the past) into the magnetic store helps optimize your data storage costs. We recommend using magnetic store writes for applications that don’t require sub-second write-to-read latency.
  • Data retention: Configure the data retention of the memory store and magnetic store to match the data ingestion (writes) throughput and query latency requirements of your application. The memory store is optimized for high throughput data writes and fast point-in-time queries. The magnetic store is optimized for long term data storage and fast analytical queries.
  • Effective query authoring: Creating queries with essential measures and dimensions, including relevant time predicates, and using limit clauses where applicable helps lower the cost of queries. See our best practices for authoring queries for more details.
  • Scheduled queries: Creating scheduled queries for frequently accessed data, including aggregates, rollups, and other real-time analytics, helps lower data storage and query costs.

Furthermore, Amazon Timestream also offers the following features at no extra charge.

  • High Availability: Although you’re charged for only one copy of your data, Amazon Timestream ensures high availability of your write and read requests by automatically replicating data and allocating resources across at least three different Availability Zones within a single AWS Region, at no extra charge.
  • Security: All data in Amazon Timestream, whether in transit or at rest, is always encrypted using AWS Key Management Service (KMS) keys. You’re not charged for the KMS calls made by Amazon Timestream for encrypting and decrypting your data.
  • Monitoring: You are not charged for the Amazon CloudWatch Metrics emitted by Amazon Timestream.
  • Metadata queries: There are no charges for queries such as SELECT 1, SHOW DATABASES, SHOW TABLES, SHOW MEASURES, and others.

Pricing details

Writes

Writes are charged based on the amount of time-series data written by an application or a scheduled query into a table. Each write request is rounded to the nearest KB.

Queries

Queries are charged based on the amount of time-series data scanned by Amazon Timestream’s serverless distributed query engine while computing data for queries sent from your applications or from your configured scheduled queries, with a 10 MB minimum per query.

Memory store

The memory store charges are calculated based on the amount of time-series data stored in the memory store of each Amazon Timestream table.

Magnetic store

The magnetic store charges are calculated based on the amount of time-series data stored in the magnetic store of each Amazon Timestream table.

SSD store

The SSD store charges are calculated based on the amount of time-series data stored in the SSD store of each Amazon Timestream table.

Metric Price
Price per GB stored per day $0.01

Pricing Examples

Example 1: Monitoring the performance and health of a video streaming application

Consider a video streaming application running on 200 EC2 instances. You want to use Amazon Timestream to store and analyze the metrics being emitted from the application so you can understand the performance and health of your application, quickly identify anomalies, resolve issues, and discover optimization opportunities.

Each EC2 instance emits four metrics (video startup time, rebuffering ratio, video playback failures, and average frame rate) per second. To identify anomalies, you’ve set up 10 queries that run every minute to identify any unusual activity over the past 10 minutes. You’ve also built a dashboard with eight widgets that display the last six hours’ worth of data so you can effectively monitor your application. This dashboard is accessed by five users at any given time and is auto-refreshed every hour. Finally, the tables in Amazon Timestream are configured to store six hours’ worth of data in the memory store and six months’ worth of data in the magnetic store.

Assumptions:

  • You also want to track the instance name, Availability Zone, and Region of each EC2 instance.
  • The dashboard queries contain essential dimensions and measures and relevant predicates, so Amazon Timestream’s distributed query engine can aggressively prune irrelevant data and scan approximately 2% of the data accumulated over the past six hours.

Monthly cost estimate (with single-measure records):
If the data is written into Amazon Timestream using the single-measure record format, the monthly cost can be estimated as shown below:

  • Using this pricing estimator, the size of a typical time series record is 600 bytes. This record will be written as 4 rows in the Amazon Timestream table, with 3 dimensions and 1 measure per row.
  • Number of writes per second: 118. This is calculated as (600 bytes * 200 records)/1 KB writes
  • Writes cost: $152.93 per month. This is calculated as 118 writes * 60 seconds * 60 minutes * 24 hours * 30 days * $0.50 per 1 million writes.
  • Memory store cost: $62.57 per month. This is calculated as 600 bytes per record * 200 records per second * 60 seconds * 60 minutes * 6 hours * 24 hours * 30 days * $0.036 per GB-hour.
  • Magnetic store cost: $52.14 per month. This is calculated as 600 bytes per record * 200 records per second * 60 seconds * 60 minutes * 24 hours * 30 days * 6 months * $0.03 per GB-month.
  • Query cost: $56.09 per month. This is calculated as follows:
    • Alerting queries cost: $42.19 per month. This is calculated as 10 queries per minute * 10 MB per query * 60 minutes * 24 hours * 30 days * $0.01 per GB of data scanned
    • Dashboarding queries cost: $13.90 per month. This is calculated as 8 widgets * 5 users * 49.4 MB per query * 24 hours * 30 days * $0.01 per GB of data scanned. Amazon Timestream’s query engine will process approximately 2% of the data accumulated over a 6-hour period (0.2% * 600 bytes per record * 200 records per second * 60 seconds * 60 minutes * 6 hours/1 MB) which is 49.4 MB.
  • Total: $323.73 per month. See this pricing estimator for additional details.

Monthly cost estimate (with multi-measure records):
If the data is written into Amazon Timestream using the multi-measure record format, the monthly cost can be estimated as shown below:

  • Using this pricing estimator, the size of a typical time series record is 180 bytes (each record will contain 3 dimensions of type varchar and 4 measures of type double).
  • Number of writes per second: 36. This is calculated as (180 bytes * 200 records) / 1 KB writes.
  • Writes cost: $46.66 per month. This is calculated as 36 writes * 60 seconds * 60 minutes * 24 hours * 30 days * $0.50 per 1 million writes.
  • Memory store cost: $18.77 per month. This is calculated as 180 bytes per record * 200 records per second * 60 seconds * 60 minutes * 6 hours * 24 hours * 30 days * $0.036 per GB-hour.
  • Magnetic store cost: $15.64 per month. This is calculated as 180 bytes per record * 200 records per second * 60 seconds * 60 minutes * 24 hours * 30 days * 6 months * $0.03 per GB-month.
  • Query cost: $46.36 per month. This is calculated as follows:
    • Alerting queries cost: $42.19 per month. This is calculated as 10 queries per minute * 10 MB per query * 60 minutes * 24 hours * 30 days * $0.01 per GB of data scanned
    • Dashboarding queries cost: $4.17 per month. This is calculated as 8 widgets * 5 users * 14.8 MB per query * 24 hours * 30 days * $0.01 per GB of data scanned. Amazon Timestream’s query engine will process  approximately 2% of the data accumulated over a 6-hour period (0.2% * 180 bytes per record * 200 records per second * 60 seconds * 60 minutes * 6 hours/1 MB) which is 14.8 MB.
  • Total: $127.43 per month. See this pricing estimator for additional details.

In this example, using multi-measure records helps lower the monthly cost by 2.5 times (from $323.73 per month to $127.43 per month). Therefore, where applicable, it is recommended to send data to Amazon Timestream using multi-measure records.
 
 

Example 2: Using real-time analytics to detect fraudulent payments and make better business decisions

Consider a payment system that processes transactions sent from multiple point-of-sale terminals distributed across major metropolitan cities in the United States. You want to use Amazon Timestream to store and analyze the transaction data so you can detect fraudulent transactions and run real-time analytics queries to answer business questions such as identifying the busiest and least-used point-of-sale terminals per hour, the busiest hour of the day for each city, the city with most transactions per hour, and other similar questions.

The system processes approximately 100,000 transactions per minute. Each transaction stored in Amazon Timestream is 100 bytes. You’ve configured 10 queries that run every minute to detect various kinds of fraudulent payments. You’ve also created 25 queries that aggregate and slice/dice your data along various dimensions to help answer your business questions. Each of these queries processes the last hour’s data. You’ve created a dashboard to display the data generated by these queries. The dashboard contains 25 widgets, it’s refreshed every hour, and it’s typically accessed by 10 users at any given time. Finally, your memory store is configured with a two-hour data retention period, and the magnetic store is configured to have a six-month data retention period.

Assumptions:

  • You plan to use multi-measure records for writing into Amazon Timestream to optimize cost.
  • You also plan to batch data. The batch size is 100 with approximately 35% of the data being repeated per batch.
  • The dashboard queries contain essential dimensions and measures and relevant predicates, so Amazon Timestream’s distributed query engine can aggressively prune irrelevant data and scan approximately 2% of the data accumulated over the past hour.

Monthly cost estimate (without scheduled queries):
If the real-time analytics queries are run as ad-hoc or automated SQL queries (instead of being defined as scheduled queries), the monthly cost can be estimated as shown below:

  • Number of writes per batch of 100 records: 7 This is calculated as (100 bytes + 99 * 65% of non-repeated data * 100 bytes) / 1 KB per write (rounded to the nearest KB)
  • Number of writes per minute: 7,000. This is calculated as 7 writes per batch * 1,000 batches of 100 records
  • Writes cost: $151.2 per month. This is calculated as 7,000 writes * 60 minutes * 24 hours * 30 days * $0.50 per 1 million writes.
  • Memory store cost: $28.97 per month. This is calculated as 100 bytes per record * 100,000 records per minute * 60 minutes * 2 hours * 24 hours * 30 days * $0.036 per GB-hour.
  • Magnetic store cost: $72.42 per month. This is calculated as 100 bytes per record * 100,000 records per minute * 60 minutes * 24 hours * 30 days * 6 months * $0.03 per GB-month.
  • Query cost: $162.89 per month. This is calculated as:
    • Alerting queries cost: $42.19 per month. This is calculated as 10 queries per minute * 10 MB per query * 60 minutes * 24 hours * 30 days * $0.01 per GB of data scanned
    • Dashboarding queries cost: $120.7 per month. This is calculated as 25 widgets * 10 users * 68.7 MB per query * 24 hours * 30 days * $0.01 per GB of data scanned. Amazon Timestream’s query engine will process approximately 2% of the data accumulated over a 6-hour period (0.2% * 100 bytes per record * 100,000 records per minute * 60 minutes * 6 hours / 1 MB) which is 68.7 MB.
  • Total: $415.47 per month. See this pricing estimator for additional details.

Monthly cost estimate (with scheduled queries):
If the real-time analytics queries are run as defined as scheduled queries, the monthly cost can be estimated as shown below:

  • Number of writes per batch of 100 records: 7 This is calculated as (100 bytes + 99 * 65% of non-repeated data * 100 bytes) / 1 KB per write (rounded to the nearest KB)
  • Number of writes per minute: 7,000. This is calculated as 7 writes per batch * 1,000 batches of 100 records
  • Writes cost: $151.2 per month. This is calculated as 7,000 writes * 60 minutes * 24 hours * 30 days * $0.50 per 1 million writes.
  • Memory store cost: $28.97 per month. This is calculated as 100 bytes per record * 100,000 records per minute * 60 minutes * 2 hours * 24 hours * 30 days * $0.036 per GB-hour.
  • Magnetic store cost: $72.42 per month. This is calculated as 100 bytes per record * 100,000 records per minute * 60 minutes * 24 hours * 30 days * 6 months * $0.03 per GB-month.
  • Query cost: $42.89 per month. This is calculated as:
    • Alerting queries cost: $42.19 per month. This is calculated as 10 queries per minute * 10 MB per query * 60 minutes * 24 hours * 30 days * $0.01 per GB of data scanned.
    • Queries on tables populated by scheduled queries: $0.70 per month. This is calculated as 10 queries per hour (10 dashboard users) * 10 MB per query * 24 hours * 30 days * $0.01 per GB of data scanned.
  • Scheduled queries: $13.57 per month. This is calculated as:
    • Query cost: $12.07 per month. This is calculated as: 25 queries per hour (analytics queries) * 68.7 MB per query * 24 hours * 30 days * $0.01 per GB of data scanned. Amazon Timestream’s query engine will process approximately 2% of the data accumulated over a 6-hour period (0.2% * 100 bytes per record * 100,000 records per minute * 60 minutes * 6 hours / 1 MB) which is 68.7 MB.
    • Writes cost: $1.50 per month. Assuming that the number of writes made by scheduled queries will be less than 1% of the total amount of data being written into Amazon Timestream, given the data is being aggregated/down-sampled to a 6-hour granularity. 1% of 302 GB is 3 GB. 3GB * $0.50 per (1M writes or 1 GB of writes) = $1.50. See this pricing estimator for additional details.
  • Total: $310.07 per month. See this pricing estimator for additional details.
 
In this example, using scheduled queries lowers the monthly cost by 1.3 times (from $415.47 per month to $310.07 per month). Therefore, where applicable, it’s recommended to schedule queries for fast and cost-effective time-series analytics.

Additional pricing resources

AWS Pricing Calculator

Easily calculate your monthly costs with AWS

Economics Resource Center

Additional resources for switching to AWS