Improve check for duplicate tag versions of all image components (#2147)

.pre-commit-config.yaml

@@ -3,18 +3,21 @@ exclude: '(venv|\.vscode|tests/k8s_schema)' # regex
 repos:
   - repo: local
     hooks:
-      - id: verify-image-tags
-        name: Verify image tags
+      - id: check-image-tags
+        language: python
+        additional_dependencies: ["jinja2", "PyYAML"]
+        name: Ensure unique tags per image
         entry: python3
-        args: [bin/verify_image_tags.py]
-        language: system
-        files: ^.*\.(tpl|yaml)$
+        args: ["bin/show-docker-images.py", "--private-registry", "--with-houston", "-c"]
+        files: '(config.yml|config.yml.j2|generate_circleci_config.py|yaml|tpl)$'
+        pass_filenames: false
       - id: circle-config-yaml
-        name: Checks for consistency between config.yml and config.yml.j2
         language: python
-        files: '(config.yml|config.yml.j2|generate_circleci_config.py|values.yaml|metadata.yaml)$'
-        entry: bin/generate_circleci_config.py
         additional_dependencies: ["jinja2", "PyYAML"]
+        name: Checks for consistency between config.yml and config.yml.j2
+        entry: bin/generate_circleci_config.py
+        files: '(config.yml|config.yml.j2|generate_circleci_config.py|yaml|tpl)$'
+        pass_filenames: false
   - repo: https://github.com/codespell-project/codespell
     rev: v2.2.6
     hooks:

bin/show-docker-images.py

@@ -8,6 +8,8 @@
 import sys
 import json
 
+from collections import defaultdict
+
 
 def get_containers_from_spec(spec):
     """Return a list of images used in a kubernetes pod spec."""
@@ -120,6 +122,12 @@ def main():
         action="store_true",
         help="include images from houston configmap",
     )
+    parser.add_argument(
+        "-c",
+        "--check-tags",
+        action="store_true",
+        help="only check for multiple tags for the same image, do not print the images",
+    )
     args = parser.parse_args()
 
     docs = helm_template(args)
@@ -140,6 +148,20 @@ def main():
             case _:
                 pass
 
+    if args.check_tags:
+        image_tag_counts = defaultdict(int)
+        for x, y in [
+            (tag.rpartition(":")[0], tag.rpartition(":")[2]) for tag in containers
+        ]:
+            image_tag_counts[x] += 1
+        for image, count in image_tag_counts.items():
+            if count > 1:
+                print(
+                    f"ALERT: {image} has {count} different tags: {', '.join([c.rpartition(':')[-1] for c in containers if image in c])}"
+                )
+        # Exit with a return code equal to the number of images with multiple tags
+        raise SystemExit(len([y for x, y in image_tag_counts.items() if y > 1]))
+
     print_results(sorted(containers))