Here’s how to save money on Google Standard Cloud Buckets:
This script shows recommendations based on the flow chart above:
bash <(curl -sL https://gist.githubusercontent.com/kienstra/fcc4c7d06a6e4d140b7fccf8d1106ced/raw/cloud-bucket-recommendations.sh) <project id>Code language: JavaScript (javascript)Lines 2, 3, and 9 are candidates for savings:
READS_LAST_30D WRITES_LAST_30D READS_LAST_90D WRITES_LAST_90D RECOMMENDATION BUCKET SIZE_GB OBJ_COUNT REGION
0 0 0 0 Move to COLDLINE example-no-activity 10890.33 15234 us-east1
0 0 1 0 Move to NEARLINE archive-legacy 2048.13 365 us-central1
4521987 2234511 5123456 2345678 None analytics-raw 1024.55 198734 europe-west1
823456 334221 912345 345678 None analytics-processed 256.02 54321 europe-west1
345 412 412 489 None terraform-state 1.24 98 us-central1
23451 19876 24567 20123 None build-artifacts 42.88 12456 us-central1
3456789 12345 3678901 13456 None media-assets 768.90 45678 asia-east1
0 0 0 0 Move to COLDLINE tmp-staging 3.12 221 us-east1
123456 876 134567 945 None audit-exports 95.44 10987 us-central1
2345 12 2567 14 None ml-datasets 1536.77 3421 us-west1
Code language: Shell Session (shell)Let’s verify the first recommendation of "Move to COLDLINE" by looking at Metrics Explorer.
If there was no read or write in the last 90 days, that means the recommendation was correct, and we should move it to "COLDLINE".
Go to the Cloud Storage page for the instance:
Once you’re in Metrics Explorer, click the first filter, and change it to:
.*Read.*|.*Write.*
Change the time to 90d (90 days).
If there was no read or write in the last 90 days, you should see “No data is available for the selected time frame”:
There’s actually no data available, so the recommendation to move to COLDLINE is correct.
You might get to the bottom of the flowchart, and see that you’ll never read or write from the bucket.
Then, it’s best to delete the bucket, and get the maximum savings.