fix: e2e_amplify_outputs_backwards_compatibility only runs once, and often fails on network connectivity issues (#3017)

The `health_checks` workflow fails on the `Setup baseline version` due
to network connectivity issues:

https://github.com/aws-amplify/amplify-backend/actions/runs/18266505828/job/52001722478#logs

<!--
Thank you for your Pull Request! Please describe the problem this PR
fixes and a summary of the changes made.
Link to any relevant issues, code snippets, or other PRs.

For trivial changes, this template can be ignored in favor of a short
description of the changes.
-->

## Problem

<!--
Describe the issue this PR is solving
-->

There are no retries on this intermittently failing task.

## Changes

<!--
Summarize the changes introduced in this PR. This is a good place to
call out critical or potentially problematic parts of the change.
-->

Implemented retries for the task.

## Validation

<!--
Describe how changes in this PR have been validated. This may include
added or updated unit, integration and/or E2E tests, test workflow runs,
or manual verification. If manual verification is the only way changes
in this PR have been validated, you will need to write some automated
tests before this PR is ready to merge.

For changes to test infra, or non-functional changes, tests are not
always required. Instead, you should call out _why_ you think tests are
not required here.

If changes affect a GitHub workflow that is not included in the PR
checks, include a link to a passing test run of the modified workflow.
--->

The workflow run on this PR will validate the change. `run-e2e` is
applied.

## Checklist

<!--
These items must be completed before a PR is ready to be merged.
Feel free to publish a draft PR before these items are complete.
-->

- [X] If this PR includes a functional change to the runtime behavior of
the code, I have added or updated automated test coverage for this
change.
- [X] If this PR requires a change to the [Project Architecture
README](../PROJECT_ARCHITECTURE.md), I have included that update in this
PR.
- [X] If this PR requires a docs update, I have linked to that docs PR
above.
- [X] If this PR modifies E2E tests, makes changes to resource
provisioning, or makes SDK calls, I have run the PR checks with the
`run-e2e` label set.

_By submitting this pull request, I confirm that my contribution is made
under the terms of the Apache 2.0 license._

Author: iankhou

.github/actions/setup_baseline_version/action.yml

@@ -4,36 +4,86 @@ inputs:
   node_version:
     description: node version used to configure environment with
     required: true
+  max_retry_attempts:
+    description: maximum number of retry attempts for network operations
+    required: false
+    default: '5'
+  initial_wait_time:
+    description: initial wait time in seconds before first retry (increases by this amount each retry)
+    required: false
+    default: '30'
 outputs:
   baseline_dir:
     description: 'Path where baseline project directory is setup'
     value: ${{ steps.move_baseline_version.outputs.baseline_dir }}
 runs:
   using: composite
   steps:
+    - name: Setup retry function
+      shell: bash
+      run: |
+        # Create a shared retry function that can be used in other steps
+        cat > /tmp/retry_function.sh << 'EOF'
+        retry_with_backoff() {
+          local max_attempts=$1
+          local initial_wait=$2
+          shift 2
+          local command="$@"
+          
+          local attempt=1
+          local wait_time=$initial_wait
+          
+          while [ $attempt -le $max_attempts ]; do
+            echo "Attempt $attempt of $max_attempts"
+            
+            if eval "$command"; then
+              echo "Command completed successfully"
+              return 0
+            fi
+            
+            if [ $attempt -lt $max_attempts ]; then
+              echo "Command failed. Waiting ${wait_time} seconds before retry..."
+              sleep $wait_time
+              wait_time=$((wait_time + initial_wait))
+            fi
+            
+            attempt=$((attempt + 1))
+          done
+          
+          echo "All retry attempts failed"
+          return 1
+        }
+        EOF
     - name: Get baseline commit sha
       id: get_baseline_commit_sha
       shell: bash
       env:
         GH_TOKEN: ${{ github.token }}
       run: |
-        if [[ ${{ github.event_name }} == 'push' ]]; then
-          # The SHA of the most recent commit on ref before the push.
-          baseline_commit_sha="${{ github.event.before }}"
-        elif [[ ${{ github.event_name }} == 'pull_request' ]]; then
-          # The SHA of the HEAD commit on base branch.
-          baseline_commit_sha="${{ github.event.pull_request.base.sha }}"
-        elif [[ ${{ github.event_name }} == 'schedule' ]] || [[ ${{ github.event_name }} == 'workflow_dispatch' ]]; then
-          # The SHA of the parent of HEAD commit on main branch.
-          # This assumes linear history of main branch, i.e. one parent.
-          # These events have only information about HEAD commit, hence the need for lookup.
-          baseline_commit_sha=$(gh api /repos/${{ github.repository }}/commits/${{ github.sha }} | jq -r '.parents[0].sha')
-        else
-          echo Unable to determine baseline commit sha;
-          exit 1;
-        fi
-        echo baseline commit sha is $baseline_commit_sha;
-        echo "baseline_commit_sha=$baseline_commit_sha" >> "$GITHUB_OUTPUT";
+        source /tmp/retry_function.sh
+
+        get_baseline_commit() {
+          if [[ ${{ github.event_name }} == 'push' ]]; then
+            # The SHA of the most recent commit on ref before the push.
+            baseline_commit_sha="${{ github.event.before }}"
+          elif [[ ${{ github.event_name }} == 'pull_request' ]]; then
+            # The SHA of the HEAD commit on base branch.
+            baseline_commit_sha="${{ github.event.pull_request.base.sha }}"
+          elif [[ ${{ github.event_name }} == 'schedule' ]] || [[ ${{ github.event_name }} == 'workflow_dispatch' ]]; then
+            # The SHA of the parent of HEAD commit on main branch.
+            # This assumes linear history of main branch, i.e. one parent.
+            # These events have only information about HEAD commit, hence the need for lookup.
+            baseline_commit_sha=$(gh api /repos/${{ github.repository }}/commits/${{ github.sha }} | jq -r '.parents[0].sha')
+          else
+            echo "Unable to determine baseline commit sha"
+            return 1
+          fi
+        }
+
+        # Execute with retries
+        retry_with_backoff ${{ inputs.max_retry_attempts }} ${{ inputs.initial_wait_time }} get_baseline_commit
+        echo "baseline commit sha is $baseline_commit_sha"
+        echo "baseline_commit_sha=$baseline_commit_sha" >> "$GITHUB_OUTPUT"
     - name: Checkout baseline version
       uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b # version 4.1.4
       with:
@@ -44,8 +94,10 @@ runs:
     - name: Install and build baseline version
       shell: bash
       run: |
-        npm ci
-        npm run build
+        source /tmp/retry_function.sh
+
+        # Execute npm ci and build with retries
+        retry_with_backoff ${{ inputs.max_retry_attempts }} ${{ inputs.initial_wait_time }} "npm ci && npm run build"
     - name: Move baseline version
       id: move_baseline_version
       shell: bash