@robmartin33 highlighted in today's (10/18/24) TF-3 meeting that there is a list of KPIs in the Finops org site that we can use as reference to help inform our efforts on the FOCUS spec (https://www.finops.org/wg/finops-kpis/). Some of these also have a playbook (in "Related Assets") describing how to calculate these KPIs from CSPs' data (including in some cases the specifics of which fields and how to query them). This raised some potential opportunities for further discussion:
1) How can we use these KPIs as a way to prioritize our work items (starting with for 1.2)?
2) How can we use the play books attached to these KPIs to benchmark and inform our approach to FOCUS fields
Proposed Approach
This is an initial proposal for discussion based on Rob's remarks and subsequent conversation in 10/18/24 TF-3 call:
For opportunity #1:
a) Survey practitioners and have them stack rank the KPIs we have already to create a "hierarchy of needs" as part of this we could also solicit verbatim feedback on why a certain KPI is important, what decisions it drives for them/their company
b) starting from the top of the hierarchy of needs, map the KPIs to the bits of data we need in FOCUS to generate these KPIs, including the bits we already gave through 1.1, and the ones we don't yet have
c) from this we should end up with a stack ranked list of new bits of data we need in FOCUS - use this list to prioritize work items for 1.2 and beyond
For opportunity #2:
d) Going top down through the stack ranked list of KPIs, go through the playbooks for each KPI to see which items already have playbook methodologies that allow the KPIs to be generated for each of the major CSPs. Where no playbooks exist, quickly draft new ones.
e) from this, identify the KPIs that can be generated from existing CSP data, and the CSP data fields required for each
f) use this stack ranked list of "low hanging fruit" as a criteria for prio decisions for V1.2 - ie if a KPI can already be generated from existing provider exports, then we should make sure that it can be generated from FOCUS as well, otherwise FOCUS is less useful than the providers exports (at least wrt that KPI).
Description
@robmartin33 highlighted in today's (10/18/24) TF-3 meeting that there is a list of KPIs in the Finops org site that we can use as reference to help inform our efforts on the FOCUS spec (https://www.finops.org/wg/finops-kpis/). Some of these also have a playbook (in "Related Assets") describing how to calculate these KPIs from CSPs' data (including in some cases the specifics of which fields and how to query them). This raised some potential opportunities for further discussion: 1) How can we use these KPIs as a way to prioritize our work items (starting with for 1.2)?
2) How can we use the play books attached to these KPIs to benchmark and inform our approach to FOCUS fields
Proposed Approach
This is an initial proposal for discussion based on Rob's remarks and subsequent conversation in 10/18/24 TF-3 call:
For opportunity #1: a) Survey practitioners and have them stack rank the KPIs we have already to create a "hierarchy of needs" as part of this we could also solicit verbatim feedback on why a certain KPI is important, what decisions it drives for them/their company b) starting from the top of the hierarchy of needs, map the KPIs to the bits of data we need in FOCUS to generate these KPIs, including the bits we already gave through 1.1, and the ones we don't yet have c) from this we should end up with a stack ranked list of new bits of data we need in FOCUS - use this list to prioritize work items for 1.2 and beyond
For opportunity #2: d) Going top down through the stack ranked list of KPIs, go through the playbooks for each KPI to see which items already have playbook methodologies that allow the KPIs to be generated for each of the major CSPs. Where no playbooks exist, quickly draft new ones. e) from this, identify the KPIs that can be generated from existing CSP data, and the CSP data fields required for each f) use this stack ranked list of "low hanging fruit" as a criteria for prio decisions for V1.2 - ie if a KPI can already be generated from existing provider exports, then we should make sure that it can be generated from FOCUS as well, otherwise FOCUS is less useful than the providers exports (at least wrt that KPI).
GitHub Issue or Reference
specification-wide topic
Context
No response
Data Submission for Discussion
No response