Cleanup docker registry

cleanup.py

#!/usr/bin/env python3

# This script will cleanup docker registry images, that are older than 7 days (see REMOVE_OLDER_THAN_DAYS).
# Note: This will only work on v2 registry.
# Before running make sure you have:
#     1. Installed requirements by running `python3 -m pip install -r requirements.txt`
#     2. Put registry password in `.registry-password` file
#     3. Updated correct values in `Constants` section
# Then run `python3 cleanup.py` or `chmod +x cleanup.py && ./cleanup.py`
# After cleaning registry, run garbage collector on registry, for example: `docker compose exec registry sh -c "/bin/registry garbage-collect --delete-untagged /etc/docker/registry/config.yml"`

import re
import sys
import json
import logging
from datetime import datetime, timezone
from docker_registry_client import DockerRegistryClient

# Constants
DOCKER_REGISTRY_USERNAME = 'admin'
DOCKER_REGISTRY_PASSWORD_FILE = "./.registry-password"
DOCKER_REGISTRY_ADDRESS = "https://localhost:5000"
REMOVE_OLDER_THAN_DAYS = 7

# Logging
logger = logging.getLogger("Docker cleanup")
logger.setLevel(logging.DEBUG)

handler = logging.StreamHandler(sys.stdout)
handler.setLevel(logging.DEBUG)

formatter = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
handler.setFormatter(formatter)

logger.addHandler(handler)

# Setup client
password = None
with open(DOCKER_REGISTRY_PASSWORD_FILE) as f:
    password = f.read()
    password = password.strip()

logger.debug(f"Connecting to registry {DOCKER_REGISTRY_ADDRESS}")
client = DockerRegistryClient(
    DOCKER_REGISTRY_ADDRESS,
    verify_ssl=False,
    username=DOCKER_REGISTRY_USERNAME,
    password=password
)
logger.debug(f"Registry version: {client.api_version}")

# List all repositories
repositories = client.repositories()
logger.debug(f"Found: {len(repositories)} repositories")

dateRegex = re.compile(r"\.[0-9]+")
now = datetime.now(timezone.utc)

for repository in repositories:
    logger.debug(f"Gathering tags for repository {repository}")
    repo = client.repository(repository)
    tags = repo.tags()
    for tag in tags:
        data, digest = repo.manifest(tag)
        try:
            historyData = data["history"][0]["v1Compatibility"]
        except KeyError:
            logger.error(f"No history data for tag {tag}")
            continue

        try:
            historyData = json.loads(historyData)
        except json.JSONDecodeError:
            logger.error(f"Could not decode history data for tag {tag}")
            continue

        # Check if the image is older than a week
        # We are throwing away microseconds, because python do not recognize them and we don't need them
        created = dateRegex.sub("", historyData["created"])
        dt = datetime.strptime(created, "%Y-%m-%dT%H:%M:%S%z")
        dateDiff = now - dt
        if dateDiff.days > REMOVE_OLDER_THAN_DAYS:
            logger.info(f"Removing tag {tag}. {dateDiff.days} days old.")
            repo.delete_manifest(digest)

requirements.txt

docker-registry-client