Amazon Athena system tables are managed Apache Iceberg tables that make completed query records available for analysis in Amazon S3 Tables. AWS announced query_executions as the first table on November 9, 2026; teams must enable capture for an AWS account and Region and integrate the S3 Tables bucket with AWS Glue Data Catalog before querying it.

The practical difference is that teams can query and join execution records in a managed table instead of building a custom history pipeline. AWS describes system tables as an additional way to work with Athena usage data, alongside existing APIs, CloudWatch metrics, and CloudTrail logs.

Plan capture around the AWS account and Region you need, then set retention and access controls before relying on the records as operational history.

What query_executions records

The table contains one row for each completed query, whether it succeeded or failed; queries still in progress do not appear. AWS says the records include query text and execution ID, submission time, workgroup and catalog context, database and table details, the submitting identity, planning, queue and execution timings, bytes scanned, state, and error information. The dpu_count field is populated only for queries run on a capacity reservation. (AWS’s announcement)

Athena stores the table in Apache Iceberg format in an aws-athena Amazon S3 Tables bucket in the same account and Region. Its fully qualified name is "s3tablescatalog/aws-athena"."metadata"."query_executions". Athena manages the table’s schema, partitioning, compaction, and snapshot maintenance; AWS recommends filtering on partition columns to limit scans, with work_group_name as an example for this table. (Amazon Athena’s system tables guide)

Enable capture and prepare access

Capture is opt-in. In the Athena console, open Administration → System tables → Configure to enable it. The setting applies to all current and future workgroups in the selected AWS account and Region.

Before querying the table, integrate the S3 Tables bucket with AWS Glue Data Catalog. AWS describes this as a one-time step for each account and Region. Query users also need permission to run Athena queries and write to the workgroup’s query-results location, read access to the Glue catalog and S3 Tables, and Lake Formation data access. A customer-managed AWS Key Management Service (AWS KMS) key adds the relevant KMS permissions. (AWS’s setup and access requirements)

The account-and-Region boundary

One enabled account and Region is the unit of coverage: system tables include workgroups there, but AWS does not support consolidating records across accounts or Regions. A team that needs that wider view cannot get it from a single system-table deployment.

Retention, capture gaps, and costs

Records have no default expiration. Retention must be configured through S3 Tables record expiration; the minimum is one day, and AWS sets no upper limit. Once written, records are delivered exactly once and are immutable: users cannot insert, update, or delete individual rows.

Disabling system tables stops new records, and AWS does not backfill the disabled period after capture resumes. That gap matters if teams depend on the table as a history of query activity.

AWS does not charge for delivering the data. Retained S3 Tables storage and maintenance incur standard charges, as do Athena queries against the table. (AWS’s billing and lifecycle documentation)

Query text may contain personally identifiable information or proprietary logic. AWS does not mask system-table data for customers; access must be governed through IAM or Lake Formation. SSE-S3 is the default encryption. If you choose a customer-managed KMS key, AWS says it cannot be changed after table creation without dropping the table and permanently removing its existing data.