Cost Price differs in Finops 11 vs 12 Dashboard

[gajadv]

🐛 Problem

I have upgraded the Finops 11 to 12. After that I have deployee the FinOps 12 Dashboard.
Cost Price differs in Finops 11 vs 12 Dashboard

👣 Repro steps

Open Finops Dashboard 11 and 12
Compare the budgeting Under Quantify
Sumary for Billed Cost month wise.

🤔 Expected

Both Dashboard should have same datta.

📷 Screenshots

TODO: If applicable, add screenshots to help explain your problem. Remove if not applicable.

ℹ️ Additional context

I run the export again for historical data and its is getting corrected.

🙋‍♀️ Ask for the community

Should have same data in both the Dashboard

[gajadv]

[RolandKrummenacher]

Hi @gajadv . Do you only see this difference in this dashboard or in the data in general? For example by comparing the results of the following queries:

Costs_v1_0
| summarize
BilledCost = todouble(round(sum(BilledCost), 2)),
EffectiveCost = todouble(round(sum(EffectiveCost), 2)),
ContractedCost = todouble(round(sum(ContractedCost), 2)),
ListCost = todouble(round(sum(ListCost), 2))
by
ChargePeriodStart = startofmonth(ChargePeriodStart)

Costs_v1_2
| summarize
BilledCost = todouble(round(sum(BilledCost), 2)),
EffectiveCost = todouble(round(sum(EffectiveCost), 2)),
ContractedCost = todouble(round(sum(ContractedCost), 2)),
ListCost = todouble(round(sum(ListCost), 2))
by
ChargePeriodStart = startofmonth(ChargePeriodStart)

[gajadv]

I see the difference in data as well. :

For the current month, data is okay and the dashboard is okay
I am manually running the config_StartBackfillProcess and this is somehow updating the data, and on both dashboards, it has the same data for the month where this pipeline and its related job completed.

Is this the need for an upgrade? We should run this pipeline: config_StartBackfillProcess

[RolandKrummenacher]

@gajadv There is no good reason why querying Costs_v1_0 in your Hub database would return different values than Costs_v1_2. Both functions project the data from the exact same underlying tables, therefore the results should be identical under normal conditions.

If backfilling your data resolves the discrepancy that is of course fine, but it should not be necessary.

[RolandKrummenacher]

@flanakin I've found an issue in Costs_v1_2() function in the hub database. The following line is leading to duplicated records:

because:

A quick fix would be to join with this subquery, but I don't fully understand the intention of this join at all:

| lookup kind=leftouter(Services | where isnotempty(x_ResourceType) | summarize take_any(ServiceSubcategory), take_any(x_ServiceModel) by x_ResourceType | project x_ResourceType, ServiceSubcategory, x_ServiceModel) on x_ResourceType

[RolandKrummenacher]

Also a kql-lookup would be much better suited than a kql-join.

[RolandKrummenacher]

This also explains why backfilling resolved the issue for @gajadv . When the data was backfilled it landed in Costs_final_v1_2 (instead of Costs_final_1_0), where the problem does not exists since the Costs_v1_2() function does not have to make a transformation/projection on the fly.

[gajadv]

what should I use.