◇ Could this help me?
Yes. This is a genuinely useful piece of reference because it states the cost of each option rather than only the benefit, and the constraint that decides between them is a physical one: whether your machine has to stay awake. Naming the once-per-hour ceiling on the cloud option is the most practical detail in it, since that is the limit people discover after building something that needed to run every five minutes. The webhook workaround is a sensible escalation rather than a hack, because moving the trigger to an external platform decouples the schedule from the vendor's own limits. General practice, and the honest caveat is that specific limits and plan caps change, so verify the current numbers rather than trusting a figure from a video.
I have a repeatable task I want running on a schedule without me starting it. Walk me through the options in order of increasing independence: an in-session loop, a local scheduled routine, a cloud-hosted routine, and an externally triggered run through a webhook. For each, tell me what has to stay powered on, how often it can actually run, and what caps apply. Then, given how often my task genuinely needs to run and whether my machine will be on, tell me which level is the right one and what I would need to change to move up a level. Check the current published limits rather than relying on figures you already have.