Spike: Ingress improvements: KEDA autoscaler

[jaskaransarkaria]

Background

Currently, we are estimating the number of ingress controllers we need. We should implement an autoscaling strategy based on the number of active connections held by each ingress controller. We can play around with the keda options in the nginx helm chart's values file

Use the current number of pods as the minimum to begin, this will give us time to see how the autoscaler behaves without impacting current performance.

In the guide, it uses 100 active connections limit from which to spawn more controllers. This is a good baseline but examine how many active connections our current controllers and consider whether this limit should be higher or lower.

Communicate changes

• [ ] post for #cloud-platform-update
• [ ] Weeknotes item
• [ ] Show the Thing/P&A All Hands/User CoP
• [ ] Announcements channel

Questions / Assumptions

Definition of done

• [ ] readme has been updated
• [ ] user docs have been updated
• [ ] another team member has reviewed
• [ ] smoke tests are green
• [ ] prepare demo for the team

Reference

How to write good user stories

[poornima-krishnasamy]

We spiked the KEDA controller here:https://github.com/ministryofjustice/cloud-platform/issues/5217 so would it be a pre-requisite for this ticket to have KEDA controller running in our cluster?

[jaskaransarkaria]

I think there needs to be an actual keda controller and once that exists the nginx config will create the scaled object and hpa