We mainly distinguish different job types and group all tasks in the below categories. Then decide the order of categories (not jobs) when is desired to run each.

• Backups
• Maintenances
• Reporting
• Automations (triggered in some way)
• Group tasks/jobs in above categories
• Decide on a sort order (when to run each category)
• Distribute the categorie thru the time windows for batch Jobs

Then, for each category

• Sort the tasks inside the category
• Estimate run time window
• Identify the master task/job (i.e. backup for dumping jobs, logrotate for the reporting), the job that makes no sense run other tasks after it.

• For progress and partial results, use some logging
• Buffer Stdout to generate a report
• Left STDERR to be catched by cron
• Use least privileged user if possible

From 00:00 to 00:30. The master job is logrotate, which rotates logs and so archives the data.

• logwatch
• awstats
• anteater
• pflogsumm
• git report branches (master-independent)
• MASTER logrotate / binnacle DB log rotate

From 01:00 to 02:00 The master jobs are the rsnapshot set, as it backups the data and dumps generated into the archive.

• autogit (each 6 hours)
• MySQL dumps
• LDAP dumps
• Redis dumps
• MASTER bakSnapshot

From 04:00 to 06:00

• Deploy code and config
• Package updates
• MASTER Maintenances