[POC] Create a new "Obligatron" project

[AshCorr]

What does this change?

Create a new "Obligatron" project, a step function consisting of 3 lambdas.

• FindAllTeams: Fetches all of the teamIds from galaxies_teams and outputs it to the next step.
• Obligation(Obligation ID): Accepts a teamId, runs some business logic related to the obligation, outputs a ObligationLambdaOutput. Invoked once per team and runs in parallel with other obligations.
• SaveObligationResults: Accepts an array of ObligationLambdaOutput, saves the results of the previous obligation lambda to a table. This table should be able to store the history of previous results.

In theory this will let us create new obligation lambdas with ease and not have to worry about orchestrating them or saving data to the DB.

Why?

As part of Ops KR1 we've been trying to source tagging data from all AWS resources so that we can measure the departments adherance to the tagging obligation.

What I've been struggling with is how to transform the data we get into something useable. Right now we've got an ever expanding SQL query thats partially stored in code, and partially stored in Grafana. It feels quite unmaintainable and difficult to expand on.

What we want to do next with KR1 is start excluding certain resources from our queries, meaning its going to get ever more complicated. Now we could create even more views, and more functions in Postgres to help us with this but theres a few problems on this front:

• Managing plpgsql code and postgres views is very painful. Every time you want to update a view or a function you need to copy over all the old code to a new migration, and make the change you wanted. plpgsql is also hard to reason and has a lot of limitations.
• The buisiness logic for the OKR ends up being duplicated across a lot of different panels, meaning if you want to change how we calculate the Obligation you end up needing to update a lot of stuff and hope you haven't missed anything. Grafana also ends up storing code instead of it being in Github. This will probably be improved with Grafana Scenes.
• How do we build a historical view of obligations? If every obligation is a bunch of bespoke panels, views, and dashboards with no standardisation between them are we going to have to figure out how we backdate each and every dashboard?

End goal

A possible end goal of this work might be to end up with a table like this in Grafana that is populated from the data this Step Function generates

Select Team: DevX Operations

Obligation	Description	Score	Passing	Drill Down
GU-1	All eligible AWS resources must be tagged	0.8	FALSE	https://grafana/d/8ea3b5g
GU-2	No FSBP Violations	1	TRUE	https://grafana/d/8ea3b5g
GU-3	Yada yada yada	0.4	FALSE	https://grafana/d/8ea3b5g
GU-4	Words words words	0.3	TRUE	https://grafana/d/8ea3b5g
GU-5	I'm not very creative	1	TRUE	https://grafana/d/8ea3b5g

[github-actions]

Deploy build 4494 of deploy::service-catalogue to CODE

All deployment options

• Deploy build 4494 of deploy::service-catalogue to CODE
• Deploy parts of build 4494 to CODE by previewing it first
• What's on CODE right now?

---

From guardian/actions-riff-raff.

[akash1810]

I think this has been superseded by #989, and can be closed?