My guess: security.
Cloud Scheduler doesn't seem to provide any way to verify the request source. If the endpoint is discovered, users could trigger the endpoint outside schedule.
Are there any other benefits of using pub/sub?
Google Cloud Scheduler: https://cloud.google.com/scheduler/
Google Cloud Pub/Sub: https://cloud.google.com/pubsub/
Top comments (2)
I'd say decoupling, throttling, and batching are the bigger ones. Security you can handle in an endpoint by requiring a token in a header that only your triggering server knows about. Which prevents third party triggers.
The real advantages are that with a pub/sub architecture the consumer is decoupled in time from the publisher. Meaning that if the consumer goes down, it can replay any events that occurred during its downtime. So where a http request requires that the consumer and the publisher both be up at the same time, pub/sub has no such requirement. It also allows the consumer to throttle and batch incoming data. http requests must be done when they are initialized, pub/sub allows the consumer to pick tasks up as it has capacity and to throttle down the number of requests it is working on to a manageable level. Batching also allows the consumer to handle more than one event at a time.
Basically pub/sub provides greater resilience at the (potential) cost of immediacy. In CAP theorem terms it helps with the P (partition) by loosening the coupling between your services/tasks/whatevers.
brilliant answer - thorough and concise. Thank you