fastlog-io 0.3.3

loguru fast wrapping supports fastapi

FastLOG

FastLOG is a lightweight wrapper around Loguru with sensible defaults, automatic trace_id, stdlib logging interception, and a directory watcher that can forward log bursts to HTTP, Telegram, or OTLP.

Features

• Zero-config logging: from fastlog import log; log.info("hello")
• Automatic trace_id injection for application logs
• Simple file logging with rotation and retention via configure()
• Standard logging records are routed through FastLOG
• fastlog CLI tails *.log files and forwards high-severity events

Installation

uv add fastlog-io
uv add "fastlog-io[otlp]"

Or with pip:

pip install fastlog-io
pip install "fastlog-io[otlp]"

Requirements:

• Python 3.12+
• loguru

Quickstart

from fastlog import configure, log

configure(
    level="INFO",
    log_path="./logs/app.log",
    rotation="10 MB",
    retention="7 days",
)

log.info("service started")

with log.trace_ctx():
    log.bind(action="api.call").info("request done")

# or use
import logging
logger = logging.getLogger("my-app")
logger.warning("cache miss")

Example output:

2025-10-13 12:23:00.193 | INFO     | 3hAhb3OFpU7zXKWBwp | ts.<module>:13 | service started
2025-10-13 12:23:00.194 | WARNING  | 1F1FwLwm4Orok0gs0a | [my-app]ts.<module>:16 | cache miss
2025-10-13 12:23:00.194 | INFO     | WG8yiDxzO3MCmWy8CB | api.call     | request done

Watcher CLI

The fastlog CLI watches a directory of *.log files, resumes from stored byte offsets, batches matching lines, and sends them to one or more destinations.

uv tool install fastlog-io

# HTTP
uv run fastlog ./logs --endpoint "$FASTLOG_NOTIFY_ENDPOINT"

# Telegram
uv run fastlog ./logs \
  --tg-token "$FASTLOG_NOTIFY_TG_TOKEN" \
  --tg-chat-id "$FASTLOG_NOTIFY_TG_CHAT_ID"

# OTLP
uv run fastlog ./logs --otlp-endpoint "http://127.0.0.1:4318/v1/logs"

Key points:

• --min-level defaults to WARNING
• offsets are stored in .multilogwatch.state.json
• HTTP and Telegram retries default to 3
• flush() drains queued batches; close() drains and stops workers
• OTLP uses the native OpenTelemetry exporter configuration model when possible

Run uv run fastlog --help for the full option list.

OTLP Notes

Install fastlog-io[otlp] to enable OTLP export.

• --otlp-endpoint explicitly overrides the collector endpoint
• when --otlp-endpoint is unset, FastLOG falls back to OTEL_EXPORTER_OTLP_LOGS_ENDPOINT, then OTEL_EXPORTER_OTLP_ENDPOINT
• --otlp-protocol, OTEL_EXPORTER_OTLP_LOGS_PROTOCOL, or OTEL_EXPORTER_OTLP_PROTOCOL selects http, grpc, or http/protobuf
• headers, timeout, TLS, and resource attributes are delegated to the native OpenTelemetry SDK/exporter env handling

Common OTLP env vars:

Variable	Description
OTEL_EXPORTER_OTLP_LOGS_ENDPOINT	Logs-specific OTLP endpoint
OTEL_EXPORTER_OTLP_ENDPOINT	Generic OTLP endpoint fallback
OTEL_EXPORTER_OTLP_LOGS_PROTOCOL	http, grpc, or http/protobuf
OTEL_EXPORTER_OTLP_PROTOCOL	Generic protocol fallback
OTEL_EXPORTER_OTLP_LOGS_HEADERS	Logs-specific request headers
OTEL_EXPORTER_OTLP_LOGS_TIMEOUT	Logs-specific exporter timeout
OTEL_SERVICE_NAME	Fallback source for resource service.namespace
OTEL_RESOURCE_ATTRIBUTES	Extra resource attributes
RUN_ENV / ENV	Resource deployment.environment when set
VERSION	Resource service.version when set

Common Environment Variables

Variable	Default	Description
LOG_PATH	empty	File output path
LOG_LEVEL	INFO	Default log level
LOG_ROTATION	100 MB	Loguru rotation policy
FASTLOG_NOTIFY_ENDPOINT	empty	HTTP destination for watcher notifications
FASTLOG_NOTIFY_LEVEL	WARNING	Default watcher threshold
FASTLOG_NOTIFY_TIMEOUT	empty	HTTP timeout override in seconds
FASTLOG_HTTP_ATTEMPTS	empty	HTTP and Telegram retry count, default 3
FASTLOG_WINDOW_SECONDS	30	Batch window in seconds
FASTLOG_NOTIFY_MAX_BYTES	4096	Max payload size per request
FASTLOG_NOTIFY_TG_TOKEN	empty	Telegram bot token
FASTLOG_NOTIFY_TG_CHAT_ID	empty	Telegram chat id
FASTLOG_OTLP_MAX_BATCH	256	Max records per OTLP export

Development

uv run --with '.[test]' pytest -q
uv build
uv run --with dist/*.whl --with pytest pytest -q

License

MIT © 2025 OWQ