Skip to main content

Centralized logging utilities for Python services.

Project description

InstruktAI Python Logger

Centralized logging utilities for Python services.

This repo provides a shared, consistent logging contract (env vars + output format + log location) intended to keep logs highly queryable (including by AIs that only read a tail window).

Background and rationale live in docs/design.md. Publishing notes live in docs/publishing.md.

Install (editable, local workspace)

From PyPI (recommended):

pip install instruktai-python-logger

From GitHub:

pip install git+ssh://git@github.com/InstruktAI/python-logger.git

API

  • Python entrypoint: instrukt_ai_logging.configure_logging(...)
  • Logger helper: instrukt_ai_logging.get_logger(name) (named **kv logging)
  • CLI entrypoint: instrukt-ai-logs (reads recent log lines)

Example:

import logging
from instrukt_ai_logging import configure_logging

configure_logging("teleclaude")
logger = logging.getLogger("teleclaude.core")
logger.info("job_started", job_id="abc123", user_id=123)

Per-process log files (multi-producer apps)

When several processes belong to the same app (a daemon plus a watcher plus a cron, for example), each one can own its own file under the same app directory by passing source=:

configure_logging("teleclaude")                       # → teleclaude.log (default)
configure_logging("teleclaude", source="docs-watch")  # → docs-watch.log
configure_logging("teleclaude", source="cron")        # → cron.log

The .log extension is always appended; pass the bare stem. Files all live under $XDG_STATE_HOME/instrukt-ai/{app}/ (fallback ~/.local/state). The CLI merges them transparently (see below).

Environment variables (contract)

Per-app prefix model (example uses TELECLAUDE_):

  • TELECLAUDE_LOG_LEVEL
  • TELECLAUDE_THIRD_PARTY_LOG_LEVEL
  • TELECLAUDE_THIRD_PARTY_LOGGERS (comma-separated logger prefixes to spotlight, e.g. httpcore,telegram)
  • TELECLAUDE_MUTED_LOGGERS (comma-separated logger prefixes to force to WARNING+, e.g. teleclaude.cli.tui)

These four per-app variables are the entire contract. There is no log-root override variable — the location is a fixed rule (see below).

Log location (contract)

Every app's logs live at the same predictable, prose-citable location on both macOS and Linux:

  • $XDG_STATE_HOME/instrukt-ai/{app}/{app}.log (fallback ~/.local/state, per the XDG Base Directory spec)

This path is always user-writable — no installer step, no sudo, no elevated permissions. configure_logging(...) raises immediately if the log file cannot be opened; logging is an essential subsystem and a process that cannot write its logs must not start blind.

Rotation

Rotation is transparent and runs entirely in user space: configure_logging(...) idempotently ensures a per-user rotation conf plus a per-user scheduler (a launchd LaunchAgent on macOS running newsyslog, a systemd user timer on Linux running logrotate) on every process start. No /etc configuration, no sudo, and no system rotation daemon ever touches these files, so rotated/recreated logs always stay owned by the producing user. Run instrukt-ai-log-setup to check rotation-asset status manually; normal operation never requires it.

CLI

# Last 10 minutes (default), follows
instrukt-ai-logs teleclaude -f

# Last 2 hours
instrukt-ai-logs teleclaude --since 2h

# Filter with regex
instrukt-ai-logs teleclaude --since 10m --grep 'logger=teleclaude'

# Follow (tail -f style) after printing the --since window
instrukt-ai-logs teleclaude --since 10m --follow

# Restrict to specific source streams (per-process files)
instrukt-ai-logs teleclaude --since 1h --logs=cron,docs-watch

Without --logs, every *.log* file in the app directory is merged by timestamp (rotation suffixes included). With --logs=stem1,stem2, only files whose stem matches an entry are read — exact stem match, so --logs=cron picks cron.log, cron.log.0, cron.log.1 and never cron-backup.log. Follow mode reads each selected file with one polling thread per file.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

instruktai_python_logger-0.6.3.tar.gz (25.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

instruktai_python_logger-0.6.3-py3-none-any.whl (17.0 kB view details)

Uploaded Python 3

File details

Details for the file instruktai_python_logger-0.6.3.tar.gz.

File metadata

  • Download URL: instruktai_python_logger-0.6.3.tar.gz
  • Upload date:
  • Size: 25.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for instruktai_python_logger-0.6.3.tar.gz
Algorithm Hash digest
SHA256 4080c62df6abc98f09ec5681f94994ddfa7453611d77eabd1f04a4356f7c5c66
MD5 3bb4ca00d126545de005401413b05ea6
BLAKE2b-256 c2771324967f2eb5f84735e397bc7470e2cd56b0adb7303e185486d1627bcd15

See more details on using hashes here.

File details

Details for the file instruktai_python_logger-0.6.3-py3-none-any.whl.

File metadata

File hashes

Hashes for instruktai_python_logger-0.6.3-py3-none-any.whl
Algorithm Hash digest
SHA256 420f1f67b5453b0f59e5544ebd018f25978899da7a6d21c017203331df4196f1
MD5 acc539cbc007580881383cee68fa1f8c
BLAKE2b-256 8b92b8f52142ff0f8dae553cda288d7d065c28354db41075abb2e4fd28d02a00

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page