ritual/checker.py
2025-04-01 16:02:23 +03:00

305 lines
13 KiB
Python

# flake8: noqa
# pylint: disable=broad-exception-raised, raise-missing-from, too-many-arguments, redefined-outer-name
# pylance: disable=reportMissingImports, reportMissingModuleSource, reportGeneralTypeIssues
# type: ignore
import warnings
warnings.filterwarnings("ignore", category=Warning)
import re
from datetime import datetime, timedelta, timezone
import subprocess
import os
import time
import random
import sys
import pkg_resources
import requests
import json
from collections import deque
required_packages = {
'grist-api': 'latest',
'colorama': 'latest',
'requests': '2.31.0',
'urllib3': '2.0.7',
'charset-normalizer': '3.3.2'
}
installed_packages = {pkg.key: pkg.version for pkg in pkg_resources.working_set}
for package, version in required_packages.items():
if package not in installed_packages or (version != 'latest' and installed_packages[package] != version):
if version == 'latest':
subprocess.check_call([sys.executable, '-m', 'pip', 'install', package, '--break-system-packages'])
else:
subprocess.check_call([sys.executable, '-m', 'pip', 'install', f"{package}=={version}", '--break-system-packages'])
from grist_api import GristDocAPI
import colorama
import logging
import socket
STATE_FILE_PATH = '/root/node/sync_state.json'
def self_update(logger):
logger.info("Checking for updates..")
script_path = os.path.abspath(__file__)
update_url = "https://gitea.vvzvlad.xyz/vvzvlad/ritual/raw/branch/main-22aug/checker.py"
try:
response = requests.get(update_url, timeout=10)
if response.status_code == 200:
current_content = ""
with open(script_path, 'r', encoding='utf-8') as f:
current_content = f.read()
if current_content != response.text:
with open(script_path, 'w', encoding='utf-8') as f:
f.write(response.text)
logger.info("Script updated successfully, restarting")
os.execv(sys.executable, ['python3'] + sys.argv)
else:
logger.info("Script is up to date")
else:
logger.error(f"Failed to download update, status code: {response.status_code}")
except Exception as e:
logger.error(f"Update error: {str(e)}")
class GRIST:
def __init__(self, server, doc_id, api_key, logger):
self.server = server
self.doc_id = doc_id
self.api_key = api_key
self.logger = logger
self.grist = GristDocAPI(doc_id, server=server, api_key=api_key)
def table_name_convert(self, table_name):
return table_name.replace(" ", "_")
def to_timestamp(self, dtime: datetime) -> int:
if dtime.tzinfo is None:
dtime = dtime.replace(tzinfo=timezone(timedelta(hours=3)))
return int(dtime.timestamp())
def insert_row(self, data, table):
data = {key.replace(" ", "_"): value for key, value in data.items()}
row_id = self.grist.add_records(self.table_name_convert(table), [data])
return row_id
def update_column(self, row_id, column_name, value, table):
if isinstance(value, datetime):
value = self.to_timestamp(value)
column_name = column_name.replace(" ", "_")
self.grist.update_records(self.table_name_convert(table), [{ "id": row_id, column_name: value }])
def delete_row(self, row_id, table):
self.grist.delete_records(self.table_name_convert(table), [row_id])
def update(self, row_id, updates, table):
for column_name, value in updates.items():
if isinstance(value, datetime):
updates[column_name] = self.to_timestamp(value)
updates = {column_name.replace(" ", "_"): value for column_name, value in updates.items()}
self.grist.update_records(self.table_name_convert(table), [{"id": row_id, **updates}])
def fetch_table(self, table):
return self.grist.fetch_table(self.table_name_convert(table))
def find_record(self, id=None, state=None, name=None, table=None):
if table is None:
raise ValueError("Table is not specified")
table_data = self.grist.fetch_table(self.table_name_convert(table))
if id is not None:
record = [row for row in table_data if row.id == id]
return record
if state is not None and name is not None:
record = [row for row in table_data if row.State == state and row.name == name]
return record
if state is not None:
record = [row for row in table_data if row.State == state]
return record
if name is not None:
record = [row for row in table_data if row.Name == name]
return record
def find_settings(self, key, table):
table = self.fetch_table(self.table_name_convert(table))
for record in table:
if record.Setting == key:
if record.Value is None or record.Value == "":
raise ValueError(f"Setting {key} blank")
return record.Value
raise ValueError(f"Setting {key} not found")
def read_state(logger):
"""Reads the last known state (status and sync count) from the state file."""
default_state = {"last_status": "Idle", "sync_count": 0}
if not os.path.exists(STATE_FILE_PATH):
logger.info(f"State file {STATE_FILE_PATH} not found, starting with default state.")
return default_state
try:
with open(STATE_FILE_PATH, 'r', encoding='utf-8') as f:
state = json.load(f)
# Validate state structure
if "last_status" not in state or "sync_count" not in state:
logger.warning(f"Invalid state file format in {STATE_FILE_PATH}, using default state.")
return default_state
# Ensure sync_count is an integer
state["sync_count"] = int(state["sync_count"])
return state
except (json.JSONDecodeError, ValueError, IOError) as e:
logger.error(f"Error reading or parsing state file {STATE_FILE_PATH}: {e}. Using default state.")
return default_state
def write_state(state, logger):
"""Writes the current state to the state file."""
try:
with open(STATE_FILE_PATH, 'w', encoding='utf-8') as f:
json.dump(state, f, indent=4)
except IOError as e:
logger.error(f"Error writing state file {STATE_FILE_PATH}: {e}")
def clean_ansi(text):
ansi_escape = re.compile(r'\x1B(?:[@-Z\\-_]|\[[0-?]*[ -/]*[@-~])')
return ansi_escape.sub('', text)
def format_number(number_str):
try:
number = int(number_str)
if number >= 1000:
return f"{number//1000}k"
return str(number)
except (ValueError, TypeError):
return "NaN" # Or some other indicator of invalid input
def check_logs(logger):
state = read_state(logger)
original_sync_count = state["sync_count"]
previous_status = state["last_status"]
try:
logs = subprocess.run(['docker', 'logs', '--since', '10m', 'infernet-node'], capture_output=True, text=True, check=True)
log_content = clean_ansi(logs.stdout)
last_checking_info = None
last_ignored_id = None
last_head_sub_id = None
# Regex patterns
checking_pattern = re.compile(r'Checking subscriptions.*last_sub_id=(\d+).*head_sub_id=(\d+).*num_subs_to_sync=(\d+)')
ignored_pattern = re.compile(r'Ignored subscription creation.*id=(\d+)')
head_sub_pattern = re.compile(r'head sub id is:\s*(\d+)')
# Use deque to efficiently get the last few relevant lines if needed,
# but processing all lines and keeping the last match is simpler here.
for line in log_content.splitlines():
match = checking_pattern.search(line)
if match:
last_checking_info = {
"last_sub_id": match.group(1),
"head_sub_id": match.group(2),
"num_subs_to_sync": int(match.group(3))
}
continue # Prioritize checking_info
match = ignored_pattern.search(line)
if match:
last_ignored_id = match.group(1)
continue
match = head_sub_pattern.search(line)
if match:
last_head_sub_id = match.group(1)
# No continue here, allows checking_info from same timeframe to override
current_status_type = "Idle"
status_message = "Idle"
if last_checking_info:
formatted_id = format_number(last_checking_info["last_sub_id"])
if last_checking_info["num_subs_to_sync"] > 0:
current_status_type = "Sync"
status_message = f"Sync: {formatted_id} ({state['sync_count']})"
logger.info(f"Node is syncing. Last sub ID: {last_checking_info['last_sub_id']}, Num subs to sync: {last_checking_info['num_subs_to_sync']}")
else:
current_status_type = "OK"
# Increment count only on transition from Sync to OK based on the *last* status type
if previous_status == "Sync":
state["sync_count"] += 1
logger.info(f"Sync completed. Sync count incremented to {state['sync_count']}.")
status_message = f"OK: {formatted_id} ({state['sync_count']})"
logger.info(f"Node is OK. Last sub ID: {last_checking_info['last_sub_id']}")
elif last_ignored_id:
# Fallback to "Ignored" logs if "Checking" is missing
formatted_id = format_number(last_ignored_id)
current_status_type = "Sync" # Assume sync if we only see ignored creations recently
status_message = f"Sync: {formatted_id} ({state['sync_count']})"
logger.info(f"Node possibly syncing (based on ignored logs). Last ignored ID: {last_ignored_id}")
elif last_head_sub_id:
# Fallback to "head sub id" if others are missing
formatted_id = format_number(last_head_sub_id)
current_status_type = "OK" # Assume OK if this is the latest relevant info
status_message = f"OK: {formatted_id} ({state['sync_count']})"
logger.info(f"Node status based on head sub id. Head sub ID: {last_head_sub_id}")
else:
logger.info("No relevant subscription log entries found in the last 10 minutes. Status: Idle.")
status_message = "Idle"
current_status_type = "Idle"
# Update state and write back if changed
state["last_status"] = current_status_type
if state["sync_count"] != original_sync_count or state["last_status"] != previous_status:
write_state(state, logger)
return {"status": status_message}
except subprocess.CalledProcessError as e:
logger.error(f"Error running docker logs command: {e.stderr or e.stdout or e}")
# Don't update state file on command error, keep previous state
return {"status": f"Error: Docker logs failed ({e.returncode})"}
except Exception as e:
logger.error(f"Unexpected error processing logs: {e}", exc_info=True)
# Don't update state file on unexpected error
return {"status": f"Error: Log processing failed"}
if __name__ == "__main__":
colorama.init(autoreset=True)
logger = logging.getLogger("Checker")
logger.setLevel(logging.INFO)
formatter = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
ch = logging.StreamHandler()
ch.setFormatter(formatter)
logger.addHandler(ch)
logger.info("Checker started")
self_update(logger)
#random_sleep = random.randint(1, 60)
#logger.info(f"Sleeping for {random_sleep} seconds")
#time.sleep(random_sleep)
grist_data = {}
with open('/root/node/grist.json', 'r', encoding='utf-8') as f:
grist_data = json.loads(f.read())
GRIST_ROW_NAME = socket.gethostname()
NODES_TABLE = "Nodes"
grist = GRIST(grist_data.get('grist_server'), grist_data.get('grist_doc_id'), grist_data.get('grist_api_key'), logger)
current_vm = grist.find_record(name=GRIST_ROW_NAME, table=NODES_TABLE)[0]
def grist_callback(msg): grist.update(current_vm.id, msg, NODES_TABLE)
for attempt in range(3):
try:
result = check_logs(logger)
grist_callback({ "Health": result["status"] })
logger.info(f"Status: {result['status']}")
break
except Exception as e:
logger.error(f"Error on attempt {attempt+1}/3: {e}")
if attempt == 2:
grist_callback({ "Health": f"Error: {e}" })
if attempt < 2:
time.sleep(5)