AWS CUR - Savings Plan Fee Discrepancy

Question

I am observing a $60 diff for Nov between the actual AWS Invoice and AWS CUR billing Data for AWS Compute Savings Plan. Here is the query on top the CUR data. 
```
select 'Savings Plan Usage' as Metrics, SUM(savings_plan_savings_plan_effective_cost) as savings_plan_savings_plan_effective_cost
from sl_aws_cur where year = '2023' and month = '11' -- Actual Usage $1200
UNION ALL
select 'Savings Plan Charge' as Metrics, SUM(line_item_unblended_cost) as savingsPlanFee
from sl_aws_cur where year = '2023' and month = '11'
AND line_item_line_item_type = 'SavingsPlanRecurringFee' -- Fee $1260 for 720 hours
```
Ideally the `savings_plan_savings_plan_effective_cost` & `SavingsPlanRecurringFee` should match exactly and same is the case for me for last couple of months. Is there any exception to that ? Thanks!

Accepted Answer

Hi Santosh, it seems likely that your Savings Plans utilization for the month of November, 2023 is less than 100%, whereas in previous months the Savings Plans may have been fully utilized.  This would indicate less compute usage in November compared to previous months.  I recommend running this query from Well-Architected labs site to check your utilization % for both November and a previous month to confirm this.  https://wellarchitectedlabs.com/cost-optimization/cur_queries/queries/aws_cost_management/#savings-plans-utilization
If you run that, it might be best to change the date filter of WHERE    CAST("concat"("year", '-', "month", '-01') AS date) = "date_trunc"('month', current_date) - INTERVAL  '1' MONTH which would produce a report for last month from current date to where year = '2023' and month = '11 like you have in your query.

Answer

Santosh - Try this query to identify your EC2 usage by account, resource_id, instance type, pricing term.  It should provide enough detail to identify which instances had reduced usage.
```
-- COMPUTE USAGE PER INSTANCE AND SPLITTING BY PURCHASE OPTION
-- For ${date_filter}, options are listed here: https://wellarchitectedlabs.com/cost/300_labs/300_cur_queries/query_help/#filtering-by-date
-- Example:
--      year = '2022' AND month = '07'
--                 -- OR --
--      line_item_usage_start_date >= now() - INTERVAL '3' month

SELECT "line_item_usage_account_id" AS "Account ID",
       "line_item_resource_id" AS "Instance ID",
       if ("product_instance_type" != '', "product_instance_type", split_part("line_item_line_item_description",' ',1)) AS "Instance Type",
       if(split_part("product_instance_type",'.',2) = 'metal', "product_instance_type",if ("product_instance_type_family" != '', "product_instance_type_family", if ("product_instance_type" != '', split_part("product_instance_type",'.',1), split_part("line_item_line_item_description",'.',1)))) AS Family,
       sum(if ("pricing_term" = 'Reserved',"line_item_usage_amount",0)) AS Reserved,
       sum(if ("pricing_term" = 'OnDemand',"line_item_usage_amount",0)) AS OnDemand,
       sum(if ("pricing_term" = '',"line_item_usage_amount",0)) AS Spot
FROM
--  ${table_name} -- Replace ${table_name} with your CUR table name
WHERE "product_product_name" = 'Amazon Elastic Compute Cloud'
      AND "line_item_resource_id" LIKE 'i-%'
      AND "line_item_operation" LIKE 'RunInstance%'
      AND (("line_item_usage_type" LIKE '%BoxUsage%') OR ("line_item_usage_type" LIKE '%SpotUsage%'))
      AND
--  ${date_filter} -- Replace ${date_filter} with an option from above.
GROUP BY "line_item_usage_account_id", "line_item_resource_id", "product_instance_type", "line_item_usage_type", "product_instance_type_family", "line_item_line_item_description"```

Answer

Here's an additional query you can run to check the $ gap between the Savings Plan commitment and usage:
```select line_item_usage_type, product_purchase_option, SUM(savings_plan_total_commitment_to_date) as SP_Total_Commitment, SUM(line_item_unblended_cost) as SP_Unblended_Cost, SUM(savings_plan_used_commitment) as SP_Used_Commitment from  where line_item_line_item_type = 'SavingsPlanRecurringFee' group by 1,2```

AWS CUR - Savings Plan Fee Discrepancy

Relevant content