AWS Releases Bedrock Invocation Logging Data Lake Solution
Understand the story
On November 11, 2026, the AWS Machine Learning Blog published a solution for an Amazon Bedrock invocation observability data lake. Designed for multi-account scenarios, the solution uses CloudWatch subscription filters to aggregate Bedrock model invocation logs from each linked account into a central account; after the logs are cleaned by Lambda, Amazon Data Firehose converts them to Parquet format and writes them to S3, while Athena provides unified SQL query capability, enabling centralized cross-account querying of every model invocation. The report is first-hand content from the official AWS blog and is currently the only coverage of this event, which is at the solution-release stage.
AI overview of reporting · Updated 3 h前
Reporting timeline
Explore different aspects of the story through its reporting.
- AWS Machine Learning BlogAmazon Bedrock tokenomics data lake
AWS has introduced an observability data lake solution for Amazon Bedrock invocations, using CloudWatch subscription filters to aggregate model invocation logs from linked accounts into a central account, cleaning them via Lambda, converting them to Parquet with Amazon Data Firehose for writing to S3, and providing unified SQL queries through Athena.
Story heat trend
Current heat9·Comparable peak10（11/11 02:00）·Comparable change over 24 hours –
The trend compares the same continuously observed sources, which may cover fewer sources than the current heat score. Hover or select the chart to inspect each hour; use the left and right arrow keys to navigate.