forked from mtusi2026/mtusi2026lab6
Собрать наблюдаемый стенд, чтобы причину деградации API и worker было видно по метрикам, логам и трассам.
This commit is contained in:
@@ -0,0 +1,98 @@
|
||||
import time
|
||||
from contextlib import contextmanager
|
||||
from pathlib import Path
|
||||
|
||||
import psycopg
|
||||
from psycopg.rows import dict_row
|
||||
from psycopg.types.json import Jsonb
|
||||
from psycopg_pool import ConnectionPool, PoolTimeout
|
||||
|
||||
from app import telemetry
|
||||
from app.errors import AppError, unavailable
|
||||
from app.faults import db_fail, db_slow_seconds
|
||||
from app.settings import settings
|
||||
|
||||
pool = None
|
||||
|
||||
|
||||
def init():
|
||||
global pool
|
||||
pool = ConnectionPool(
|
||||
settings.database_url,
|
||||
min_size=1,
|
||||
max_size=settings.pool_max,
|
||||
timeout=2,
|
||||
max_waiting=16,
|
||||
check=ConnectionPool.check_connection,
|
||||
kwargs={
|
||||
"row_factory": dict_row,
|
||||
"autocommit": False,
|
||||
"connect_timeout": 3,
|
||||
"options": "-c timezone=UTC -c statement_timeout=4000",
|
||||
},
|
||||
open=True,
|
||||
reconnect_timeout=2,
|
||||
)
|
||||
|
||||
|
||||
def migrate():
|
||||
script = Path(__file__).with_name("schema.sql").read_text(encoding="utf-8")
|
||||
statements = []
|
||||
buffer = []
|
||||
for line in script.splitlines():
|
||||
stripped = line.strip()
|
||||
if not stripped or stripped.startswith("--"):
|
||||
continue
|
||||
buffer.append(line)
|
||||
if stripped.endswith(";"):
|
||||
statements.append("\n".join(buffer))
|
||||
buffer = []
|
||||
with pool.connection() as conn:
|
||||
conn.execute("SELECT pg_advisory_lock(861006)")
|
||||
for statement in statements:
|
||||
conn.execute(statement)
|
||||
conn.execute("SELECT pg_advisory_unlock(861006)")
|
||||
conn.commit()
|
||||
|
||||
|
||||
def jsonb(value):
|
||||
return Jsonb(value)
|
||||
|
||||
|
||||
@contextmanager
|
||||
def connection():
|
||||
if db_fail():
|
||||
raise unavailable()
|
||||
delay = db_slow_seconds()
|
||||
started = time.perf_counter()
|
||||
span = telemetry.tracer().start_as_current_span(
|
||||
"db.transaction",
|
||||
attributes={"db.system": "postgresql"},
|
||||
)
|
||||
span.__enter__()
|
||||
try:
|
||||
if delay:
|
||||
time.sleep(delay)
|
||||
with pool.connection() as conn:
|
||||
try:
|
||||
yield conn
|
||||
conn.commit()
|
||||
except Exception:
|
||||
conn.rollback()
|
||||
raise
|
||||
except AppError:
|
||||
raise
|
||||
except (PoolTimeout, psycopg.OperationalError, psycopg.InterfaceError) as exc:
|
||||
raise unavailable() from exc
|
||||
finally:
|
||||
span.__exit__(None, None, None)
|
||||
telemetry.record_db(time.perf_counter() - started)
|
||||
|
||||
|
||||
def stats():
|
||||
if pool is None:
|
||||
return 0, 0
|
||||
data = pool.get_stats()
|
||||
used = int(data.get("pool_size", 0)) - int(data.get("pool_available", 0))
|
||||
waiting = int(data.get("requests_waiting", 0))
|
||||
return max(used, 0), waiting
|
||||
Reference in New Issue
Block a user