A separate crontab for every customer, automatically

Your SaaS lets each customer schedule their own report generation. The naive design: one shared scheduler, customer IDs in the job payload, fingers crossed nobody starves anyone else.

The Hoody design: each customer gets their own container. Each container has its own hoody-cron service. Customer A schedules a 9am daily run; customer B schedules */5 * * * *. They live on different crontabs, different filesystems, different process trees, on the same bare metal box.

When customer B's job hangs for 40 minutes, customer A's 9am run still fires on time. There's no shared queue, no priority lane, no noisy-neighbor incident report. The isolation is the substrate, not a feature you wrote.

One customer's cron can't starve another's because they aren't on the same crontab.