aws-samples / aws-samples/amazon-redshift-config-compare
Missing Analysis Output
- Dominant language
- Python
- Stars
- 25
- Forks
- 10
- PR merge metrics
- No merged PRs in 30d
Description
Hello,
We've launched the CF template and ran the SF. However, at the end of the process the performance metrics are _not_ available to query. We're currently just testing the tool and rerunning a handful of queries with the simple replay log configuration. The cluster has audit_logging and user_activity enabled. We do not have a ddl_and_copy and sql_script configured. We are not overwriting the replay or extraction configuration.
At the moment we're using it within the same account as the initial Redshift Cluster and Snapshot - but this is a staging account. For production we will use it across account.
Having a look at the logs we can see a few things suspect things,
From the lambda function we can see an unload to (s3://bucket/data/raw_results) and the table it's trying to unload (public.redshift_config_comparison_raw). However, there is no data in S3, and there's no data for the table in Redshift.
Stepping back another layer, we noticed in the batch logs there was a line that stated:`Analysis not enabled for this replay.`
From what I can see the `replay_bootstrap.sh` does not set an `analysis_output` or `analysis_iam_role` for the SimpleReplay config, which ultimately results in the`Analysis not enabled for replay` line. Should Analysis be enabled?
We've been following the [blog](https://aws.amazon.com/blogs/big-data/compare-different-node-types-for-your-workload-using-amazon-redshift/) and readme, but there's nothing (we can see) about enabling analysis. It sounds like it _should_ be the expected output.
I'm wondering if you could give us some pointers to where we should begin to dive into this.
Contributor guide
Assessment
This issue has not been assessed yet.