In #698 I suggested changing the retention to 30 days for the raw fastly logs, as the successful request data logs will contain all the same information for successful tiles, so the only loss is failed requests and non-tile requests, which we've never needed.
Since the log aggregation has been running for a couple weeks without issues, I recommend turning on a bucket policy to delete files >30 days old. This will bring the bucket size down from 86TB to about 4TB.
In #698 I suggested changing the retention to 30 days for the raw fastly logs, as the successful request data logs will contain all the same information for successful tiles, so the only loss is failed requests and non-tile requests, which we've never needed.
Since the log aggregation has been running for a couple weeks without issues, I recommend turning on a bucket policy to delete files >30 days old. This will bring the bucket size down from 86TB to about 4TB.