Auto DevOps ignoring .gitlab/auto-deploy-values.yaml file values?

Summary

Readiness probe continues to go to port 5000 after being overridden with a .gitlab/auto-deploy-values.yaml file instructing it to test port 3000.

Steps to reproduce

1. Create simple node.js app that listens with Express on port 3000.
2. Create Dockerfile for app (no expose or publish).
3. Commit to project monitored by gitlab auto devops.
4. Observe failure of probe.
5. Add .gitlab/auto-deploy-values.yaml with a service section containing internalPort: 3000 and externalPort: 3000 directives in that section and commit.
6. Observe failure of probe again, note that the failure is still referencing http://10.x.y.z:5000/ in the kubectl get events output.

Example Project

Unable to provide at this time. Basic framework contains:

Dockerfile

FROM node:14
WORKDIR /usr/src/app

COPY package*.json ./
RUN npm install
COPY . .
CMD [ "npm", "start" ]

.gitlab/auto-deploy-values.yaml

deploymentApiVersion: apps/v1
service:
  internalPort: 3000
  externalPort: 3000

What is the current bug behavior?

Readiness probe fails because it tests the wrong port.

What is the expected correct behavior?

Readiness probe should succeed.

Relevant logs and/or screenshots

Output from kubectl get events -n .... states:

Readiness probe failed: Get "http://10.244.4.70:5000/": dial tcp 10.244.4.70:5000: connect: connection refused