stella-logger 1.0.1

Strict and safe logging SDK.

Stella Logger

Strict and safe structured logging SDK for Python. Stella Logger centralizes log event definitions, validates payloads with Pydantic 2.x, renders messages from templates (no ad-hoc overrides), and adapts output for multiple clouds via pluggable “output modes.”

• Centralized definitions: LogDefinition + LogRegistry keep event keys/codes unique.
• Validated payloads: Optional Pydantic models ensure schema correctness.
• Template-based messages: Messages are rendered from message_template; callers cannot inject arbitrary text.
• Cloud-ready output: Standard logging backend with modes for JSON line (CloudWatch), nested extra (GCP), or flat extra (Lambda JSON).

Why

Long-running services accumulate ad-hoc logs that drift. Stella Logger treats logs as events with stable keys/codes, enforceable schemas, and predictable output formats, keeping observability consistent across teams and clouds.

Install

pip install stella-logger

Quickstart

import logging
from pydantic import BaseModel
from stella_logger.core import StellaCoreLogger, StellaCoreSettings, StellaSeverity
from stella_logger.schema import LogDefinition, LogRegistry, LogKind
from stella_logger.logger import StellaLogger

class DbQueryTimeoutPayload(BaseModel):
    host: str
    port: int
    elapsed_ms: int | None = None

definitions = [
    LogDefinition(
        key="DB_QUERY_TIMEOUT",
        numeric=1053,
        message_template="DB timeout on {host}:{port}, elapsed={elapsed_ms}ms",
        severity=StellaSeverity.ERROR,
        kind=LogKind.ERROR,
        category="TECH_ERROR",
        schema_model=DbQueryTimeoutPayload,
    )
]

registry = LogRegistry(definitions)
core = StellaCoreLogger(
    settings=StellaCoreSettings(
        service_name="my-app",
        service_version="1.5.17",
        structured_message=True,  # JSON_MESSAGE mode
    ),
    base_logger=logging.getLogger("myapp.stella"),
)
logger = StellaLogger(core=core, registry=registry)

logger.log_error("DB_QUERY_TIMEOUT", extra={"host": "db1", "port": 5432, "elapsed_ms": 1200})

Output modes (core)

• JSON_MESSAGE: msg=json.dumps(payload); best for AWS CloudWatch Logs (non-Lambda).
• FLAT_EXTRA: msg=payload["message"]; extra is flattened (message dropped); good for Lambda JSON logging.
• NESTED_EXTRA: msg=payload["message"]; extra nests payload under payload_attr_name (e.g., json_fields for GCP).

Project layout

• SDK code: src/stella_logger/
• Tests: tests/
• Docs (design/spec/recipes): docs/

詳細な設計思想・仕様・クラウド別レシピは docs/ を参照してください。

Development

• Install deps: poetry install --with dev
• Run tests: poetry run pytest
• Lint: poetry run ruff check

License

MIT