new file: crowdsec/Dockerfile

new file:   default-environment.env
	new file:   docker-compose.yml
	new file:   docker-exporter/Dockerfile
	new file:   docker-exporter/Dockerfile.old
	new file:   docker-exporter/TEST_EXAMPLE.md
	new file:   docker-exporter/exporter.py
	new file:   docker-exporter/exporter.py.old
	new file:   services-up.sh
	new file:   telegraf.conf
	new file:   traefik/certs/.gitkeep
	new file:   traefik/dynamic.yml
	new file:   traefik/letsencrypt/acme.json
	new file:   traefik/traefik.yml
This commit is contained in:
2026-04-14 08:09:33 +10:00
parent 1141f452a8
commit bb00e8684d
14 changed files with 1421 additions and 0 deletions
+6
View File
@@ -0,0 +1,6 @@
FROM crowdsecurity/crowdsec:latest
COPY config/ /etc/crowdsec
# Install required components at build time
RUN cscli hub update && \
cscli collections install crowdsecurity/traefik --force
+18
View File
@@ -0,0 +1,18 @@
PROJECT_ROOT=/home/raspi/raspi
REMOTE_PROJECT_ROOT=/home/nixos/raspi
DOMAIN=lan.ddnsgeek.com
TZ=Australia/Brisbane
EMAIL=wayne.bennett@live.com
# Docker Socket Proxy
DOCKER_SOCKET_PROXY_LOG_LEVEL=info
DOCKER_SOCKET_PROXY_HOST=tcp://docker-socket-proxy:2375
# Docker Update Exporter
DOCKER_EXPORTER_LOG_LEVEL=INFO
# Crowdsec
CROWDSEC_COLLECTIONS=crowdsecurity/traefik
+179
View File
@@ -0,0 +1,179 @@
version: "3.9"
services:
traefik:
profiles: ["all"]
image: traefik:3
container_name: traefik
restart: always
depends_on:
- docker-socket-proxy
# environment:
# TRAEFIK_PROVIDERS_DOCKER_ENDPOINT: ${DOCKER_SOCKET_PROXY_HOST}
ports:
- "80:80"
- "443:443"
- "8080:8080"
volumes:
# - /var/run/docker.sock:/var/run/docker.sock:ro
- ${PROJECT_ROOT}/traefik/letsencrypt:/letsencrypt
- ${PROJECT_ROOT}/traefik/traefik.yml:/etc/traefik/traefik.yml:ro
- ${PROJECT_ROOT}/traefik/dynamic.yml:/etc/traefik/dynamic.yml:ro
networks:
- proxy
uptime-kuma:
profiles: ["all"]
image: louislam/uptime-kuma
container_name: uptime-kuma
restart: always
depends_on:
- traefik
- docker-socket-proxy
environment:
DOCKER_HOST: ${DOCKER_SOCKET_PROXY_HOST}
volumes:
- ${PROJECT_ROOT}/uptime-kuma:/app/data
# - /var/run/docker.sock:/var/run/docker.sock
networks:
- proxy
labels:
- traefik.enable=true
# Router
- traefik.http.routers.kuma.rule=Host(`kuma.lan.ddnsgeek.com`)
- traefik.http.routers.kuma.entrypoints=websecure
- traefik.http.routers.kuma.tls=true
- traefik.http.routers.kuma.tls.certresolver=le
# Service -> container port
- traefik.http.services.kuma.loadbalancer.server.port=3001
# - traefik.http.routers.kuma.middlewares=default-chain@file
# ports:
# - 8080:3001
portainer-agent:
profiles: ["all"]
image: portainer/agent
restart: always
depends_on:
- docker-socket-proxy
ports:
- 9001:9001
container_name: portainer-agent
environment:
DOCKER_HOST: ${DOCKER_SOCKET_PROXY_HOST}
volumes:
# - /var/run/docker.sock:/var/run/docker.sock
- /var/lib/docker/volumes:/var/lib/docker/volumes
- /:/host
networks:
- proxy
crowdsec:
profiles: ["all"]
image: crowdsecurity/crowdsec:latest
# build: ./crowdsec
container_name: crowdsec
restart: always
environment:
- COLLECTIONS=${CROWDSEC_COLLECTIONS}
volumes:
- ${PROJECT_ROOT}/data/logs:/logs:ro
- ${PROJECT_ROOT}/crowdsec/data:/var/lib/crowdsec/data
- ${PROJECT_ROOT}/crowdsec/config:/etc/crowdsec
networks:
- proxy
docker-update-exporter:
profiles: ["all"]
build: ${REMOTE_PROJECT_ROOT}/docker-exporter
container_name: docker-update-exporter
volumes:
# - /var/run/docker.sock:/var/run/docker.sock
- ${PROJECT_ROOT}:/compose:ro
# - ${PROJECT_ROOT}/docker-compose.yml:/compose/docker-compose.yml:ro
- ${REMOTE_PROJECT_ROOT}/docker-exporter/data:/data:rw
- ${PROJECT_ROOT}/.docker/config.json:/root/.docker/config.json:ro
environment:
LOG_LEVEL: ${DOCKER_EXPORTER_LOG_LEVEL}
DOCKER_HOST: ${DOCKER_SOCKET_PROXY_HOST}
ports:
- "9105:9105"
restart: unless-stopped
depends_on:
- docker-socket-proxy
cap_drop:
- ALL
security_opt:
- no-new-privileges:true
networks:
- proxy
healthcheck:
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:9105/metrics')"]
interval: 30s
timeout: 5s
retries: 3
start_period: 10s
telegraf:
profiles: ["monitoring","all","prometheus"]
image: telegraf:latest
# group_add:
# - "123"
# privileged: true
container_name: telegraf
restart: unless-stopped
depends_on:
- docker-socket-proxy
# cap_drop:
# - ALL
security_opt:
- no-new-privileges:true
volumes:
# - /var/run/docker.sock:/var/run/docker.sock:ro
- ${PROJECT_ROOT}/telegraf.conf:/etc/telegraf/telegraf.conf:ro
networks:
# - edge
- proxy
ports:
- 9273:9273
healthcheck:
test: ["CMD-SHELL", "curl -f http://localhost:9273/metrics || exit 1"]
interval: 30s
timeout: 5s
retries: 3
start_period: 10s
docker-socket-proxy:
profiles: ["monitoring","all","prometheus","prometheus-exporters"]
image: tecnativa/docker-socket-proxy:latest
container_name: docker-socket-proxy
restart: unless-stopped
environment:
LOG_LEVEL: ${DOCKER_SOCKET_PROXY_LOG_LEVEL}
DISTRIBUTION: 1
CONTAINERS: 1
EVENTS: 1
IMAGES: 1
INFO: 1
NETWORKS: 1
PING: 1
POST: 1
SERVICES: 1
TASKS: 1
VERSION: 1
VOLUMES: 1
volumes:
- /var/run/docker.sock:/var/run/docker.sock:ro
cap_drop:
- ALL
security_opt:
- no-new-privileges:true
networks:
- proxy
networks:
proxy:
name: proxy
+15
View File
@@ -0,0 +1,15 @@
FROM python:3.11-slim
#RUN groupadd -g 1000 appuser || true && \
# useradd -m -u 1000 -g 1000 -s /bin/bash appuser
#RUN groupadd -g 999 docker || true && usermod -aG docker appuser
WORKDIR /app
COPY exporter.py .
RUN mkdir -p /data
RUN pip install --no-cache-dir docker prometheus_client requests pyyaml
#USER appuser
CMD ["python", "-u", "exporter.py"]
+8
View File
@@ -0,0 +1,8 @@
FROM python:3.11-slim
WORKDIR /app
COPY exporter.py .
RUN pip install docker prometheus_client requests pyyaml
CMD ["python", "exporter.py"]
+24
View File
@@ -0,0 +1,24 @@
# Exporter image-mapping test example
Run the exporter in dry-run mode to print the `service -> image:tag` mapping without starting the metrics loop:
```bash
SERVICES_UP_SCRIPT=/workspace/docker/services-up.sh python monitoring/docker-exporter/exporter.py --dry-run
```
Example output excerpt:
```json
{
"crowdsec": "crowdsecurity/crowdsec:latest",
"docker-update-exporter": "python:3.11-slim",
"nextcloud-webapp": "nextcloud:production",
"node-red": "nodered/node-red:latest",
"prometheus": "prom/prometheus:latest",
"traefik": "traefik:3"
}
```
This confirms the exporter now reports images for both:
- services with explicit `image:` values, and
- services using `build:` contexts.
+528
View File
@@ -0,0 +1,528 @@
#!/usr/bin/env python3
import argparse
import os
import re
import time
import json
import logging
import docker
import yaml
from prometheus_client import Gauge, start_http_server
# --- Logging ---
LOG_LEVEL = os.getenv("LOG_LEVEL", "DEBUG").upper()
logging.basicConfig(
level=getattr(logging, LOG_LEVEL, logging.DEBUG),
format="%(asctime)s [%(levelname)s] %(message)s"
)
logger = logging.getLogger("docker-update-exporter")
# --- Config ---
EXPORTER_PORT = 9105
CHECK_INTERVAL = 3600
CACHE_TTL = int(os.getenv("CACHE_TTL", "300"))
SERVICES_UP_SCRIPT = os.getenv("SERVICES_UP_SCRIPT", "/compose/services-up.sh")
CACHE_FILE = os.getenv("CACHE_FILE", "/data/remote_digest_cache.json")
DRY_RUN = os.getenv("DRY_RUN", "false").lower() in ("1", "true", "yes")
try:
client = docker.from_env()
except Exception as e:
logger.warning(f"Docker client unavailable at startup: {e}")
client = None
# --- Metrics ---
CONTAINER_UPDATE = Gauge(
"docker_container_update_available",
"1 if container image is out of date (compose drift or registry), 0 otherwise",
["container", "compose_image", "running_image", "com_docker_compose_project"]
)
LAST_CHECK = Gauge(
"docker_image_update_last_check_timestamp",
"Last time the update check ran (unix timestamp)"
)
def set_container_update_metric(container_name, compose_image, running_image, project_name, update_flag):
"""Set update metric for a container and log the emitted metric payload."""
metric_labels = {
"container": container_name,
"compose_image": compose_image or "unknown",
"running_image": running_image,
"com_docker_compose_project": project_name,
}
CONTAINER_UPDATE.labels(**metric_labels).set(update_flag)
logger.info(
"Metric emitted: docker_container_update_available=%s labels=%s",
update_flag,
metric_labels,
)
def set_last_check_metric():
"""Set and log the timestamp for the most recent check cycle."""
ts = time.time()
LAST_CHECK.set(ts)
logger.info("Metric emitted: docker_image_update_last_check_timestamp=%s", ts)
# --- Persistent Cache ---
def load_cache():
if not os.path.exists(CACHE_FILE):
logger.info(f"Cache file does not exist yet: {CACHE_FILE}")
return {}
try:
with open(CACHE_FILE, "r") as f:
cache = json.load(f)
logger.info(f"Loaded {len(cache)} cached remote digests")
return cache
except Exception as e:
logger.error(f"Failed to load cache: {e}")
return {}
def save_cache():
try:
os.makedirs(os.path.dirname(CACHE_FILE), exist_ok=True)
with open(CACHE_FILE, "w") as f:
json.dump(REMOTE_DIGEST_CACHE, f)
logger.debug(f"Saved {len(REMOTE_DIGEST_CACHE)} remote digests to cache")
except Exception as e:
logger.error(f"Failed to save cache: {e}")
REMOTE_DIGEST_CACHE = load_cache()
now = time.time()
REMOTE_DIGEST_CACHE = {
image: (digest, ts)
for image, (digest, ts) in REMOTE_DIGEST_CACHE.items()
if now - ts < CACHE_TTL
}
# --- Helpers ---
def get_project_prefix_from_script(script_path):
prefix = "core-"
if not os.path.exists(script_path):
return prefix
try:
with open(script_path) as f:
for line in f:
m = re.match(r'PROJECT\s*=\s*["\']?([^"\']+)', line)
if m:
return m.group(1) + "-"
except Exception as e:
logger.warning(f"Failed reading project prefix: {e}")
return prefix
def get_local_digest(image_name):
"""
Return the local digest for the specific image reference.
"""
if client is None:
return None
try:
img = client.images.get(image_name)
digests = img.attrs.get("RepoDigests", [])
logger.debug(f"RepoDigests for {image_name}: {digests}")
for entry in digests:
if "@" in entry:
digest = entry.split("@", 1)[1]
logger.debug(f"Local digest for {image_name}: {digest}")
return digest
logger.debug(f"No RepoDigest found for {image_name}")
except Exception as e:
logger.debug(f"Could not get local digest for {image_name}: {e}")
return None
def get_remote_digest(image_name):
"""
Return the upstream digest for the exact platform-specific image that Docker
would pull on this host. This avoids false positives with multi-arch images
where the registry manifest-list digest differs from the pulled image digest.
"""
now = time.time()
cached = REMOTE_DIGEST_CACHE.get(image_name)
if cached:
digest, ts = cached
if now - ts < CACHE_TTL:
logger.debug(f"Using cached remote digest for {image_name}: {digest}")
return digest
if client is None:
return None
try:
registry_data = client.images.get_registry_data(image_name)
digest = None
# docker SDK versions differ; try the common fields in order
if hasattr(registry_data, "id") and registry_data.id:
digest = registry_data.id
elif hasattr(registry_data, "attrs"):
digest = (
registry_data.attrs.get("Descriptor", {}).get("digest")
or registry_data.attrs.get("digest")
)
if digest:
REMOTE_DIGEST_CACHE[image_name] = (digest, now)
save_cache()
logger.debug(f"Remote digest for {image_name}: {digest}")
return digest
logger.warning(f"No remote digest found for {image_name}")
return None
except Exception as e:
logger.debug(f"Error fetching remote digest for {image_name}: {e}")
return None
# --- Dockerfile Image Extraction ---
def parse_dockerfile_for_image(dockerfile_path):
if not os.path.exists(dockerfile_path):
return None
try:
arg_defaults = {}
last_from = None
with open(dockerfile_path) as df:
for line in df:
line = line.strip()
if not line or line.startswith("#"):
continue
if line.upper().startswith("ARG "):
arg_body = line[4:].strip()
if "=" in arg_body:
key, value = arg_body.split("=", 1)
arg_defaults[key.strip()] = value.strip()
continue
# Prefer LABEL with image if present.
if "LABEL" in line and "image=" in line:
match = re.search(r'image=["\']?([^"\']+)["\']?', line)
if match:
image_name = normalize_image_name(substitute_dockerfile_args(match.group(1), arg_defaults))
logger.debug(f"Found LABEL image={image_name} in {dockerfile_path}")
return image_name
if line.upper().startswith("FROM "):
from_clause = line[5:].strip()
if from_clause.startswith("--"):
split_clause = from_clause.split(None, 1)
if len(split_clause) < 2:
continue
from_clause = split_clause[1]
parts = from_clause.split()
if not parts:
continue
candidate = substitute_dockerfile_args(parts[0], arg_defaults)
if candidate and candidate.lower() != "scratch":
last_from = normalize_image_name(candidate)
if last_from:
logger.debug(f"Found base FROM {last_from} in {dockerfile_path}")
return last_from
except Exception as e:
logger.debug(f"Error reading Dockerfile {dockerfile_path}: {e}")
return None
def normalize_image_name(image_name):
if not image_name:
return None
if "@" in image_name:
return image_name
if ":" in image_name.rsplit("/", 1)[-1]:
return image_name
return f"{image_name}:latest"
def is_compose_build_placeholder(image_name, project_name):
if not image_name:
return False
candidate = str(image_name)
project_prefix = f"{project_name}-"
if candidate.startswith(project_prefix):
return True
# Keep backward-compatible behavior for historical default project prefix.
return candidate.startswith("core-")
def substitute_dockerfile_args(value, arg_defaults):
if not value:
return value
pattern = re.compile(r"\$\{([^}]+)\}|\$([A-Za-z_][A-Za-z0-9_]*)")
def replacer(match):
expr = match.group(1)
simple = match.group(2)
if simple:
return arg_defaults.get(simple, "")
if ":-" in expr:
var_name, default_value = expr.split(":-", 1)
return arg_defaults.get(var_name, default_value)
if "-" in expr:
var_name, default_value = expr.split("-", 1)
return arg_defaults.get(var_name, default_value)
return arg_defaults.get(expr, "")
return pattern.sub(replacer, value)
def expand_compose_path(path_value, project_root):
raw = str(path_value)
raw = raw.replace("${PROJECT_ROOT}", project_root).replace("$PROJECT_ROOT", project_root)
return os.path.expandvars(raw)
def get_project_root_from_script(script_path):
if not script_path:
return os.getcwd()
return os.path.dirname(os.path.abspath(script_path))
# --- Compose parsing ---
def get_compose_files_from_script(script_path):
files = []
if not os.path.exists(script_path):
return files
base_dir = get_project_root_from_script(script_path)
try:
with open(script_path) as f:
content = f.read()
match = re.search(r'FILES\s*=\s*\((.*?)\)', content, re.DOTALL)
if match:
for line in match.group(1).splitlines():
line = line.strip()
if line.startswith("-f"):
path = line[2:].strip()
if path:
full = os.path.normpath(os.path.join(base_dir, path))
files.append(full)
except Exception as e:
logger.warning(f"Failed parsing services-up.sh: {e}")
return files
def parse_project_name_from_script(script_path):
project = "core"
if not os.path.exists(script_path):
return project
try:
with open(script_path) as f:
for line in f:
m = re.match(r'PROJECT\s*=\s*["\']?([^"\']+)', line)
if m:
project = m.group(1)
break
except Exception as e:
logger.warning(f"Failed reading project name: {e}")
return project
def resolve_local_build_image(service_name, project_name):
if client is None:
return None
try:
images = client.images.list(filters={"label": f"com.docker.compose.service={service_name}"})
for image in images:
labels = image.attrs.get("Config", {}).get("Labels", {}) or {}
if labels.get("com.docker.compose.project") != project_name:
continue
for tag in image.tags:
if tag and "<none>" not in tag:
logger.debug(f"Resolved local compose image for {service_name}: {tag}")
return normalize_image_name(tag)
except Exception as e:
logger.debug(f"Could not inspect local build metadata for {service_name}: {e}")
return None
def parse_compose_services(compose_files, project_name, project_root):
svc_map = {}
for f in compose_files:
if not os.path.exists(f):
logger.warning(f"Compose file from services-up.sh is missing: {f}")
continue
try:
with open(f) as stream:
data = yaml.safe_load(stream) or {}
for svc_name, svc_def in data.get("services", {}).items():
image = normalize_image_name(svc_def.get("image"))
profiles = svc_def.get("profiles", [])
build_ctx = svc_def.get("build")
dockerfile_path = None
from_dockerfile = None
local_built_image = None
if build_ctx:
if isinstance(build_ctx, dict):
context = build_ctx.get("context", ".")
dockerfile = build_ctx.get("dockerfile", "Dockerfile")
else:
context = build_ctx
dockerfile = "Dockerfile"
compose_dir = os.path.dirname(f)
context_expanded = expand_compose_path(context, project_root)
if os.path.isabs(context_expanded):
context_path = context_expanded
else:
context_path = os.path.normpath(os.path.join(compose_dir, context_expanded))
dockerfile_expanded = expand_compose_path(dockerfile, project_root)
dockerfile_path = os.path.normpath(os.path.join(context_path, dockerfile_expanded))
from_dockerfile = normalize_image_name(parse_dockerfile_for_image(dockerfile_path))
local_built_image = resolve_local_build_image(svc_name, project_name)
placeholder_image = is_compose_build_placeholder(image, project_name) or is_compose_build_placeholder(local_built_image, project_name)
if placeholder_image:
resolved_image = from_dockerfile or image or local_built_image or f"{project_name}-{svc_name}:latest"
else:
resolved_image = image or local_built_image or from_dockerfile or f"{project_name}-{svc_name}:latest"
svc_map[svc_name] = {
"image": resolved_image,
"profiles": profiles,
"build_context": build_ctx,
"compose_file": f,
"dockerfile": dockerfile_path
}
except Exception as e:
logger.warning(f"Failed parsing {f}: {e}")
logger.debug(f"Service image mapping: {svc_map}")
return svc_map
# --- Main check ---
def check_containers():
if client is None:
logger.error("Docker client is unavailable; skipping check cycle")
return
set_last_check_metric()
CONTAINER_UPDATE.clear()
project_name = parse_project_name_from_script(SERVICES_UP_SCRIPT)
project_root = get_project_root_from_script(SERVICES_UP_SCRIPT)
compose_files = get_compose_files_from_script(SERVICES_UP_SCRIPT)
svc_map = parse_compose_services(compose_files, project_name, project_root)
containers = client.containers.list()
pending_metrics = []
remote_targets = set()
for container in containers:
proj = container.labels.get("com.docker.compose.project")
if not proj:
continue
svc = container.labels.get("com.docker.compose.service")
running = container.attrs["Config"]["Image"]
compose_image = None
if svc in svc_map:
compose_image = svc_map[svc]["image"]
local_digest = get_local_digest(running)
remote_target = compose_image or running
# If we cannot determine a local digest, we cannot compare and should
# avoid spending a registry lookup for this container.
if local_digest:
remote_targets.add(remote_target)
pending_metrics.append({
"container_name": container.name,
"service": svc,
"compose_image": compose_image,
"running_image": running,
"project_name": proj,
"remote_target": remote_target,
"local_digest": local_digest,
})
remote_digests = {target: get_remote_digest(target) for target in remote_targets}
for payload in pending_metrics:
local_digest = payload["local_digest"]
remote_target = payload["remote_target"]
remote_digest = remote_digests.get(remote_target)
update_flag = 1 if (local_digest and remote_digest and local_digest != remote_digest) else 0
logger.info(
"Digest comparison: container=%s service=%s running=%s target=%s local=%s remote=%s",
payload["container_name"],
payload["service"],
payload["running_image"],
remote_target,
local_digest,
remote_digest,
)
set_container_update_metric(
container_name=payload["container_name"],
compose_image=payload["compose_image"],
running_image=payload["running_image"],
project_name=payload["project_name"],
update_flag=update_flag,
)
def dump_service_image_mapping():
project_name = parse_project_name_from_script(SERVICES_UP_SCRIPT)
project_root = get_project_root_from_script(SERVICES_UP_SCRIPT)
compose_files = get_compose_files_from_script(SERVICES_UP_SCRIPT)
svc_map = parse_compose_services(compose_files, project_name, project_root)
mapping = {name: data["image"] for name, data in sorted(svc_map.items())}
logger.info("Service to image mapping:")
logger.info(json.dumps(mapping, indent=2, sort_keys=True))
return mapping
# --- Runner ---
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Docker image update exporter")
parser.add_argument("--dry-run", action="store_true", help="Only print service->image mapping and exit")
parser.add_argument(
"--services-up-script",
default=SERVICES_UP_SCRIPT,
help=f"Path to services-up script (default: {SERVICES_UP_SCRIPT})",
)
parser.add_argument(
"--cache-file",
default=CACHE_FILE,
help=f"Path to digest cache file (default: {CACHE_FILE})",
)
parser.add_argument(
"--log-level",
default=LOG_LEVEL,
help=f"Logging level (default: {LOG_LEVEL})",
)
args = parser.parse_args()
effective_log_level = str(args.log_level).upper()
logging.getLogger().setLevel(getattr(logging, effective_log_level, logging.DEBUG))
logger.setLevel(getattr(logging, effective_log_level, logging.DEBUG))
SERVICES_UP_SCRIPT = args.services_up_script
CACHE_FILE = args.cache_file
REMOTE_DIGEST_CACHE = load_cache()
now = time.time()
REMOTE_DIGEST_CACHE = {
image: (digest, ts)
for image, (digest, ts) in REMOTE_DIGEST_CACHE.items()
if now - ts < CACHE_TTL
}
if DRY_RUN or args.dry_run:
dump_service_image_mapping()
raise SystemExit(0)
start_http_server(EXPORTER_PORT)
while True:
try:
check_containers()
except Exception as e:
logger.exception(f"update check failed: {e}")
time.sleep(CHECK_INTERVAL)
+514
View File
@@ -0,0 +1,514 @@
#!/usr/bin/env python3
import os
import re
import time
import json
import logging
import docker
import requests
import yaml
from prometheus_client import Gauge, start_http_server
# --- Logging ---
LOG_LEVEL = os.getenv("LOG_LEVEL", "INFO").upper()
logging.basicConfig(
level=getattr(logging, LOG_LEVEL, logging.INFO),
format="%(asctime)s [%(levelname)s] %(message)s"
)
logger = logging.getLogger("docker-update-exporter")
# --- Config ---
EXPORTER_PORT = 9105
CHECK_INTERVAL = 60
CACHE_TTL = 6 * 3600
SERVICES_UP_SCRIPT = "/compose/services-up.sh"
CACHE_FILE = "/data/remote_digest_cache.json"
client = docker.from_env()
# --- Metrics ---
CONTAINER_UPDATE = Gauge(
"docker_container_update_available",
"1 if container image is out of date (compose drift or registry), 0 otherwise",
["container", "compose_image", "running_image", "com_docker_compose_project"]
)
LAST_CHECK = Gauge(
"docker_image_update_last_check_timestamp",
"Last time the update check ran (unix timestamp)"
)
# --- Persistent Cache ---
def load_cache():
if not os.path.exists(CACHE_FILE):
logger.info(f"Cache file does not exist yet: {CACHE_FILE}")
return {}
try:
with open(CACHE_FILE, "r") as f:
cache = json.load(f)
logger.info(f"Loaded {len(cache)} cached remote digests")
logger.debug(f"Cache contents: {cache}")
return cache
except Exception as e:
logger.error(f"Failed to load cache from {CACHE_FILE}: {e}")
return {}
def save_cache():
try:
os.makedirs(os.path.dirname(CACHE_FILE), exist_ok=True)
with open(CACHE_FILE, "w") as f:
json.dump(REMOTE_DIGEST_CACHE, f)
logger.debug(
f"Saved {len(REMOTE_DIGEST_CACHE)} entries to cache file {CACHE_FILE}"
)
except Exception as e:
logger.error(f"Failed to save cache to {CACHE_FILE}: {e}")
REMOTE_DIGEST_CACHE = load_cache()
# --- Helpers ---
def get_project_prefix_from_script(script_path):
project_prefix = "core-" # fallback
if not os.path.exists(script_path):
logger.warning(
f"services-up script not found at {script_path}, using fallback project prefix {project_prefix}"
)
return project_prefix
try:
with open(script_path, "r") as f:
for line in f:
line = line.strip()
m = re.match(r'PROJECT\s*=\s*["\']?([^"\']+)["\']?', line)
if m:
project_prefix = m.group(1) + "-"
logger.debug(
f"Detected compose project prefix from script: {project_prefix}"
)
break
except Exception as e:
logger.error(f"Failed reading project prefix from {script_path}: {e}")
return project_prefix
def get_local_digest(image_name):
try:
img = client.images.get(image_name)
digests = img.attrs.get("RepoDigests", [])
logger.debug(f"Local RepoDigests for {image_name}: {digests}")
if digests:
digest = digests[0].split("@")[1]
logger.debug(f"Local digest for {image_name}: {digest}")
return digest
logger.info(f"No local digest found for image {image_name}")
except Exception as e:
logger.warning(f"Failed to retrieve local digest for {image_name}: {e}")
return None
def get_remote_digest(image_name):
now = time.time()
original = image_name
# Cache hit
if original in REMOTE_DIGEST_CACHE:
digest, ts = REMOTE_DIGEST_CACHE[original]
age = now - ts
if age < CACHE_TTL:
logger.debug(
f"Using cached remote digest for {original} "
f"(age={int(age)}s, ttl={CACHE_TTL}s): {digest}"
)
return digest
logger.info(
f"Cache entry expired for {original} "
f"(age={int(age)}s > ttl={CACHE_TTL}s)"
)
try:
if "/" not in image_name:
registry = "docker.io"
repo = "library/" + image_name
else:
parts = image_name.split("/")
if "." in parts[0] or ":" in parts[0]:
registry = parts[0]
repo = "/".join(parts[1:])
else:
registry = "docker.io"
repo = image_name
if ":" in repo:
repo, tag = repo.rsplit(":", 1)
else:
tag = "latest"
logger.debug(
f"Resolving remote digest for {original}: "
f"registry={registry}, repo={repo}, tag={tag}"
)
token = None
manifest_url = None
if registry in ["docker.io", "registry-1.docker.io"]:
logger.debug(f"Requesting Docker Hub token for {repo}")
token_res = requests.get(
"https://auth.docker.io/token",
params={
"service": "registry.docker.io",
"scope": f"repository:{repo}:pull"
},
timeout=10
)
logger.debug(
f"Docker Hub token response for {repo}: "
f"status={token_res.status_code}"
)
token = token_res.json().get("token")
manifest_url = (
f"https://registry-1.docker.io/v2/{repo}/manifests/{tag}"
)
elif registry == "ghcr.io":
logger.debug(f"Requesting GHCR token for {repo}")
token_res = requests.get(
"https://ghcr.io/token",
params={
"service": "ghcr.io",
"scope": f"repository:{repo}:pull"
},
timeout=10
)
logger.debug(
f"GHCR token response for {repo}: "
f"status={token_res.status_code}"
)
token = token_res.json().get("token")
manifest_url = f"https://ghcr.io/v2/{repo}/manifests/{tag}"
else:
logger.warning(
f"Unsupported registry '{registry}' for image {original}"
)
return None
if not token:
logger.warning(
f"No authentication token returned for {original}"
)
return None
logger.debug(f"Requesting manifest for {original}: {manifest_url}")
res = requests.get(
manifest_url,
headers={
"Authorization": f"Bearer {token}",
"Accept": "application/vnd.docker.distribution.manifest.v2+json"
},
timeout=10
)
logger.debug(
f"Manifest response for {original}: "
f"status={res.status_code}"
)
if res.status_code == 200:
digest = res.headers.get("Docker-Content-Digest")
logger.info(
f"Fetched remote digest for {original}: {digest}"
)
REMOTE_DIGEST_CACHE[original] = (digest, now)
save_cache()
logger.debug(
f"Cached remote digest for {original}: {digest}"
)
return digest
if res.status_code == 429:
logger.warning(
f"Registry rate limit hit while fetching {original}"
)
elif res.status_code in [401, 403]:
logger.warning(
f"Authentication failed while fetching {original}: "
f"status={res.status_code}"
)
else:
logger.warning(
f"Unexpected manifest response for {original}: "
f"status={res.status_code}, body={res.text[:250]}"
)
except Exception as e:
logger.error(f"Failed to fetch remote digest for {original}: {e}")
return None
def get_compose_files_from_script(script_path):
files = []
if not os.path.exists(script_path):
logger.error(f"services-up script not found: {script_path}")
return files
base_dir = os.path.dirname(script_path)
try:
with open(script_path, "r") as f:
content = f.read()
match = re.search(r'FILES\s*=\s*\((.*?)\)', content, re.DOTALL)
if not match:
logger.warning(
f"No FILES=(...) block found in {script_path}"
)
return files
lines = match.group(1).splitlines()
for line in lines:
line = line.strip()
if line.startswith("-f"):
rel_path = line[2:].strip()
if rel_path:
full_path = os.path.normpath(
os.path.join(base_dir, rel_path)
)
logger.debug(
f"Resolved compose file: {rel_path} -> {full_path}"
)
files.append(full_path)
logger.info(f"Found {len(files)} compose files")
except Exception as e:
logger.error(f"Failed parsing compose files from {script_path}: {e}")
return files
def parse_compose_files(compose_files):
service_to_image = {}
for f in compose_files:
if not os.path.exists(f):
logger.warning(f"Compose file missing: {f}")
continue
try:
with open(f, "r") as stream:
data = yaml.safe_load(stream) or {}
services = data.get("services", {})
logger.debug(
f"Parsing {len(services)} services from compose file {f}"
)
for service_name, service_def in services.items():
image = service_def.get("image")
is_built = False
if not image:
is_built = True
build_ctx = service_def.get("build")
logger.debug(
f"Service {service_name} is build-based, build config={build_ctx}"
)
if isinstance(build_ctx, dict):
context_path = build_ctx.get("context", ".")
dockerfile_path = os.path.join(
context_path,
build_ctx.get("dockerfile", "Dockerfile")
)
elif isinstance(build_ctx, str):
context_path = build_ctx
dockerfile_path = os.path.join(
context_path, "Dockerfile"
)
else:
dockerfile_path = None
if dockerfile_path and os.path.exists(dockerfile_path):
try:
with open(dockerfile_path, "r") as df:
for line in df:
line = line.strip()
if (
line.upper().startswith("LABEL")
and "image=" in line
):
m = re.search(
r'image=["\']?([^"\']+)["\']?',
line
)
if m:
image = m.group(1)
logger.debug(
f"Found upstream image label for {service_name}: {image}"
)
break
except Exception as e:
logger.warning(
f"Failed reading Dockerfile {dockerfile_path}: {e}"
)
if not image:
image = f"{service_name}:latest"
logger.info(
f"No image label found for build service {service_name}, "
f"defaulting to {image}"
)
service_to_image[service_name] = (image, is_built)
except Exception as e:
logger.error(f"Failed parsing compose file {f}: {e}")
logger.info(f"Mapped {len(service_to_image)} compose services to images")
logger.debug(f"Service/image mapping: {service_to_image}")
return service_to_image
def check_containers():
logger.info("Starting container update check")
CONTAINER_UPDATE.clear()
project_prefix = get_project_prefix_from_script(SERVICES_UP_SCRIPT)
compose_files = get_compose_files_from_script(SERVICES_UP_SCRIPT)
service_to_image = parse_compose_files(compose_files)
containers = client.containers.list()
logger.info(f"Checking {len(containers)} running containers")
for container in containers:
project_label = container.labels.get("com.docker.compose.project")
if not project_label:
logger.debug(
f"Skipping non-compose container {container.name}"
)
continue
service_label = container.labels.get("com.docker.compose.service")
running_image = container.attrs["Config"]["Image"]
logger.debug(
f"Evaluating container={container.name}, "
f"service={service_label}, project={project_label}, "
f"running_image={running_image}"
)
compose_image = None
is_built = False
if service_label and service_label in service_to_image:
compose_image, is_built = service_to_image[service_label]
if is_built:
compose_image_name, _, _ = compose_image.partition(":")
compose_image = f"{project_prefix}{compose_image_name}"
update_flag = 0
if is_built:
if running_image != compose_image:
logger.info(
f"Update detected for build-based container {container.name}: "
f"running image {running_image} != expected {compose_image}"
)
update_flag = 1
else:
local_digest = get_local_digest(running_image)
remote_digest = get_remote_digest(
service_to_image[service_label][0]
)
if local_digest and remote_digest and local_digest != remote_digest:
logger.info(
f"Remote image update available for {container.name}: "
f"{local_digest} != {remote_digest}"
)
update_flag = 1
else:
if running_image != compose_image:
logger.info(
f"Compose drift detected for {container.name}: "
f"running image {running_image} != compose image {compose_image}"
)
update_flag = 1
else:
local_digest = get_local_digest(running_image)
remote_digest = get_remote_digest(running_image)
if local_digest and remote_digest and local_digest != remote_digest:
logger.info(
f"Registry update available for {container.name}: "
f"{local_digest} != {remote_digest}"
)
update_flag = 1
CONTAINER_UPDATE.labels(
container=container.name,
compose_image=compose_image if compose_image else "unknown",
running_image=running_image,
com_docker_compose_project=project_label
).set(update_flag)
logger.info(
f"Container {container.name}: "
f"running={running_image}, "
f"compose={compose_image}, "
f"update_available={update_flag}"
)
LAST_CHECK.set(time.time())
logger.info("Container update check complete")
if __name__ == "__main__":
logger.info(
f"Starting Docker update exporter on port {EXPORTER_PORT} "
f"with LOG_LEVEL={LOG_LEVEL}"
)
start_http_server(EXPORTER_PORT)
while True:
try:
check_containers()
except Exception as e:
logger.exception(f"Unhandled error during update check: {e}")
time.sleep(CHECK_INTERVAL)
Executable
+10
View File
@@ -0,0 +1,10 @@
#!/usr/bin/env bash
CONTEXT="raspi"
ENV="default-environment.env"
PROJECT="raspi"
FILES=(
-f docker-compose.yml
)
docker --context $CONTEXT compose -p $PROJECT --env-file $ENV "${FILES[@]}" $1 $2 $3 $4 $5 $6 $7 $8 $9
+9
View File
@@ -0,0 +1,9 @@
[agent]
interval = "10s"
[[inputs.docker]]
endpoint = "tcp://docker-socket-proxy:2375"
gather_services = false
[[outputs.prometheus_client]]
listen = ":9273"
View File
+27
View File
@@ -0,0 +1,27 @@
http:
middlewares:
crowdsec:
plugin:
crowdsec-bouncer:
crowdsecMode: live
crowdsecLapiKey: HeneLa2mazFVzl5+DQRKOdchBuJxKdjrHsHBE/03Acs
crowdsecLapiHost: crowdsec:8080
crowdsecLapiScheme: http
secHeaders:
headers:
browserXssFilter: true
contentTypeNosniff: true
frameDeny: true
# sslRedirect: true
#HSTS Configuration
stsIncludeSubdomains: true
stsPreload: true
stsSeconds: 15552000
forceSTSHeader: true
customFrameOptionsValue: "SAMEORIGIN"
default-chain:
chain:
middlewares:
- secHeaders
- crowdsec
File diff suppressed because one or more lines are too long
+58
View File
@@ -0,0 +1,58 @@
api:
dashboard: true
insecure: true # keep false if exposing via router
ping: {}
log:
level: INFO
accessLog:
filePath: /logs/access.log
format: json
entryPoints:
web:
address: ":80"
http:
redirections:
entryPoint:
to: websecure
scheme: https
websecure:
address: ":443"
http:
middlewares:
- default-chain@file
providers:
docker:
endpoint: "tcp://docker-socket-proxy:2375"
exposedByDefault: false
file:
filename: /etc/traefik/dynamic.yml
watch: true
experimental:
plugins:
crowdsec-bouncer:
moduleName: github.com/maxlerebourg/crowdsec-bouncer-traefik-plugin
version: v1.4.2
certificatesResolvers:
le:
acme:
email: wayne.bennett@live.com
storage: /letsencrypt/acme.json
httpChallenge:
entryPoint: web
metrics:
prometheus:
# entryPoint: metrics:9100 # optional, default is "metrics"
buckets:
- 0.1
- 0.3
- 1.2
- 5.0
addEntryPointsLabels: true # add labels for each entrypoint
addServicesLabels: true # add labels for each service