A reliability monitor for Linux, in the spirit of the Windows Reliability Monitor. It reads the systemd journal and coredumps into a local SQLite database and shows a 30-day score chart with a grouped timeline of crashes, kernel/hardware errors (OOM, GPU, disk, thermal), failed services and unclean shutdowns.
Works on any systemd distribution. No distro-specific code.
- Daily score chart with legend and summary, per-category and per-day filtering (with a visible day banner), search with "show more", live counts per category
- Undo toast when hiding an event; shortcuts: Ctrl+F filter, F5 / Ctrl+R refresh
- Journal context (±30 s) for each event, copy as a Markdown bug report
- Mute noisy events, per-app crash trends, per-kernel score comparison
- Desktop notifications for new panics and crashes (background timer)
- Headless text/Markdown report, light and dark themes, keyboard and screen-reader support
Needs Python 3, PyGObject, GTK 4 and libadwaita (e.g. python3-gi, gir1.2-adw-1).
Your user must be allowed to read the journal (group systemd-journal).
python3 app.py # GUI
python3 collector.py [--follow] # collect once / keep collecting
python3 collector.py --report --md # last 7 days as Markdown (also: app.py --report, --days N)
The setup menu in the app can enable background collection (a systemd user timer) and an app-menu entry.
for t in tests/test_*.py; do PYTHONPATH=. python3 $t; done
core.py is pure logic with no GTK imports, so it can be ported on its own.
GPL-3.0-or-later. See LICENSE.
