Daily System Report
# From dashboard overload to a single morning email
Starting Point
Every morning, checking various dashboards, apps, and tools — SSL status, backups, server uptime, logs, cloud costs. Takes time, and things get missed.
There was no unified system that queries all relevant sources, prioritizes intelligently, and summarizes everything in a single, clean report.
Implementation
Definition of the most important data sources: SSL certificates, backups, system logs, uptime, cloud costs, calendar, tasks.
Modular collector pipeline: Each collector is a standalone Bash script that delivers JSON metrics. New sources are integrated in minutes.
Python-based report generator with Jinja2 templates: takes the JSON data, prioritizes by severity, and renders a responsive HTML template.
Automatic delivery via SMTP at 07:00 through a systemd timer. Discord webhook planned as a second channel.
Functions
Modular Bash collectors deliver JSON. SSL, backup, logs, uptime, costs — each independent, easily extensible.
Critical issues on top, routine at the bottom. Overdue detection, color-coded severity levels, clear visual hierarchy.
Historical data in MariaDB. CPU/RAM/disk trends over weeks, anomaly detection through baseline comparison.
Email as primary channel, Discord webhook as supplement. Configurable per recipient and severity.
Technical Details
The architecture follows a clear principle: Collect, Process, Present. Each layer is independently replaceable.
Bash is available everywhere, JSON is machine-readable and easy to parse. Each collector is a standalone script that doesn't block the rest on failure.
Python for data processing and logic, Jinja2 for clean HTML templates. Separation of logic and presentation enables quick template changes without code modifications.
systemd timers offer logging, dependencies, and error handling out of the box. No silent failures like with cron.
Status
Over 180 automated reports in 3 months of operation. 12 anomalies detected early, zero unnoticed outages.
The system has established itself as a reliable daily companion — a single glance at the inbox is enough to know the state of all systems.
grep -c returns exit code 1 when count is 0 — || echo 0 produces "0\n0". Fix: var=$(cmd | grep -c) || true
Jinja2's dict.items conflicts with Python dict method — use backups['items'] instead of backups.items in templates.