fix: Resolve cache timeout issues in CI/CD pipelines

- Add fail-on-cache-miss: false to all cache actions in workflows
- Create improved runner configuration (config_cache_fixed.yaml) with:
  - Fixed cache host: host.docker.internal
  - Fixed cache port: 44029
  - Host network mode for better container networking
- Add cache troubleshooting scripts:
  - fix-cache-issues.sh (Linux/macOS)
  - fix-cache-issues.ps1 (Windows)
- Update all workflows: api-gateway, frontend, service-adapters, api-docs, ci

This resolves the 'connect ETIMEDOUT 172.31.0.3:44029' errors by:
1. Making cache failures non-fatal
2. Using proper Docker networking configuration
3. Providing tools to diagnose and fix cache issues

.gitea/workflows/api-docs.yml

@@ -43,6 +43,7 @@ jobs:
           ${{ runner.os }}-node-${{ matrix.node-version }}-
           ${{ runner.os }}-node-
           ${{ runner.os }}-
+        fail-on-cache-miss: false
       id: npm-cache
       
     - name: Cache status
@@ -136,6 +137,7 @@ jobs:
         restore-keys: |
           ${{ runner.os }}-node-18-
           ${{ runner.os }}-node-
+        fail-on-cache-miss: false
         
     - name: Install dependencies
       run: |