[INF-7324] (opt-in) hpa for jse deployment

[lukefoster11]

Easy to enable hpa for the js-executor deployment. Scale on 70% mem and 70% cpu usage.
Ask from Colgate

[greptile-apps]

Confidence Score: 5/5

The PR appears safe to merge.

No blocking failure remains.

Important Files Changed

Filename	Overview
charts/retool/templates/deployment_js_executor.yaml	Adds the gated JS executor HPA, matching-request validation, and conditional omission of Deployment replicas.
charts/retool/values.yaml	Adds the chart’s public JS executor autoscaling defaults and configuration guidance.
values.yaml	Keeps the repository-level values copy synchronized with the deployable chart defaults.
charts/retool/ci/test-js-executor-hpa-option.yaml	Adds a fixture covering HPA rendering, CPU and memory targets, replica ownership, and scaling behavior.
charts/retool/Chart.yaml	Increments the chart patch version for the new opt-in feature.

_{Reviews (6): Last reviewed commit: "simplify" | Re-trigger Greptile}

[lukefoster11]

@greptile review me again

[lukefoster11]

@greptile review me again pls

[lukefoster11]

@greptile can you review again

[lukefoster11]

@greptile review me again pls

[init-b]

LGTM! just wondering what happens to JS executions when it scales down?

[lukefoster11]

LGTM! just wondering what happens to JS executions when it scales down?

The replicacount will just go down to the min replica count (1 by default but configurable). JSE shouldn't hit performance issues while cpu and mem utilization are below 70% so JS execution speed shouldn't be affected. One thing I just realized is that this HPA depends on having metrics-server enabled, so I will add a comment that if that isn't enabled then the HPA won't work properly and will stay at the min replica count.

[init-b]

LGTM! just wondering what happens to JS executions when it scales down?

The replicacount will just go down to the min replica count (1 by default but configurable).

@lukefoster11 any implications about the in-flight process in the container we kill? I really don't know how that currently works in our Cloud environment.

[lukefoster11]

LGTM! just wondering what happens to JS executions when it scales down?

The replicacount will just go down to the min replica count (1 by default but configurable).

@lukefoster11 any implications about the in-flight process in the container we kill? I really don't know how that currently works in our Cloud environment.

It's a graceful termination, same as when we terminate all jse pods when we restart a deployment. In-flight requests complete before the pod spins down

[init-b]

LGTM! just wondering what happens to JS executions when it scales down?

The replicacount will just go down to the min replica count (1 by default but configurable).

@lukefoster11 any implications about the in-flight process in the container we kill? I really don't know how that currently works in our Cloud environment.

It's a graceful termination, same as when we terminate all jse pods when we restart a deployment. In-flight requests complete before the pod spins down

Thanks for the context Luke!