AWS redshift documentation change
Summary
Updated documentation to reflect RG provisioned clusters handling Spectrum queries differently. Added details about column semantics changes and deprecated columns for RG clusters.
Security assessment
Changes focus on monitoring view behavior differences between cluster types without addressing vulnerabilities, security controls, or threats. Modifications explain technical implementation details of query metrics collection.
Diff
diff --git a/redshift/latest/dg/c-spectrum-metrics.md b/redshift/latest/dg/c-spectrum-metrics.md index ea2db8209..aa7284161 100644 --- a//redshift/latest/dg/c-spectrum-metrics.md +++ b//redshift/latest/dg/c-spectrum-metrics.md @@ -6,0 +7,2 @@ +Note about RG provisioned clusters + @@ -11 +13 @@ Amazon Redshift will no longer support the creation of new Python UDFs starting -This topic describes system views that you can use to monitor Redshift Spectrum queries. +This topic describes system views that you can use to monitor data lake queries. @@ -13 +15 @@ This topic describes system views that you can use to monitor Redshift Spectrum -You can monitor Amazon Redshift Spectrum queries using the following system views: +You can monitor data lake queries using the following system views: @@ -17 +19 @@ You can monitor Amazon Redshift Spectrum queries using the following system view -Use the SVL_S3QUERY view to get details about Redshift Spectrum queries (S3 queries) at the segment and node slice level. +Use the SVL_S3QUERY view to get details about data lake queries at the segment and node slice level. @@ -21 +23 @@ Use the SVL_S3QUERY view to get details about Redshift Spectrum queries (S3 quer -Use the SVL_S3QUERY_SUMMARY view to get a summary of all Amazon Redshift Spectrum queries (S3 queries) that have been run on the system. +Use the SVL_S3QUERY_SUMMARY view to get a summary of all data lake queries that have been run on the system. @@ -38,0 +41,42 @@ The following are some things to look for in SVL_S3QUERY_SUMMARY: +## Note about RG provisioned clusters + +On RG provisioned clusters, Redshift Spectrum queries run on the cluster's own compute rather than on the dedicated Spectrum fleet used by RA3 and DC2 provisioned clusters. **[SVL_S3QUERY](./r_SVL_S3QUERY.html)** and **[SVL_S3QUERY_SUMMARY](./r_SVL_S3QUERY_SUMMARY.html)** continue to be populated on RG clusters. The following columns have different semantics or are deprecated. + +### Columns populated with different semantics on RG provisioned clusters + +On RA3 and DC2 provisioned clusters, these columns describe rows, bytes, and work units moving between the cluster and the Spectrum fleet. On RG provisioned clusters, they describe the equivalent work performed by the cluster's native reader: + + * **s3_scanned_rows** — rows read directly from Amazon S3 by the cluster's native reader (pre-filter). + + * **s3_scanned_bytes** — total scan-range size in bytes processed by the cluster's native reader. + + * **s3query_returned_rows** — rows produced after filter pushdown by the cluster's native reader. + + * **s3query_returned_bytes** — bytes produced after filter pushdown by the cluster's native reader. + + * **splits** — number of scan ranges consumed by the cluster's native reader. + + * **total_split_size** — total size of all scan ranges consumed, in bytes. + + * **max_split_size** — size of the largest scan range consumed, in bytes. + + + + +### Columns deprecated on RG provisioned clusters + +These columns describe Spectrum-fleet concepts that do not exist on RG. On RG clusters they are logged as -1 in STL_S3QUERY, and therefore appear as -1 or are not meaningful in SVL_S3QUERY and SVL_S3QUERY_SUMMARY: + + * **total_retries** , **max_retries** — On RG, retries occur at the Amazon S3 client level. Use STL_S3CLIENT and STL_S3CLIENT_ERROR for retry details. + + * **max_request_duration** , **avg_request_duration** — RG does not use the Spectrum request model. + + * **max_request_parallelism** , **avg_request_parallelism** — RG does not use Spectrum request tokens. + + * **slowdown_count** , **max_concurrent_slowdown_count** — On RG, Amazon S3 slowdowns are tracked in STL_S3CLIENT. + + + + +For aggregated per-query metrics on RG clusters (partitions, files scanned, rows and bytes returned, file format, file location, listing and partition-fetch timing), you can also use the [SYS_EXTERNAL_QUERY_DETAIL](./SYS_EXTERNAL_QUERY_DETAIL.html) monitoring view. +