Daily/weekly digest email: scheduled job health summary

[csoscd]

Summary

Instead of individual alert emails per failure, send a configurable scheduled digest summarising the health of all jobs in one email.

Motivation / Use Case

Operators managing many low-priority jobs don't want inbox noise from individual alerts. A single daily or weekly summary gives an overview of job health without requiring constant attention.

Proposed Behaviour

• Configurable schedule: daily or weekly, with a configurable send time
• Each job listed with: last run timestamp, last exit code, 7-day success rate, and current failure streak (if any)
• Jobs with active failure streaks are highlighted at the top
• Only sent if there is at least one job to report on
• Respects existing mail.* config (same sender/recipient as failure alerts)

Affected Component

• Agent
• Web UI (config UI for digest settings)

First Implementation Idea

1. New scheduled script — bin/send-digest.php queries execution_log and cronjobs for aggregate health data per job and builds the email body.
2. Crontab entry — send-digest.php is added to the agent container's own crontab at the configured time (daily/weekly). The schedule is written by entrypoint.sh based on config values.
3. Config keys — notifications.digest.enabled, notifications.digest.frequency (daily/weekly), notifications.digest.send_time (HH:MM), notifications.digest.weekday (0–6, for weekly mode).
4. Email rendering — plain-text and HTML variants via MailNotifier; a simple table layout per job (name, last run, success rate, streak).
5. Web UI — digest settings added to the notifications section of the settings/housekeeping page.