Actor.call: Add option to keep retrying on not enough account memory.

[metalwarrior665]

Users often struggle with understanding that when you run an Actor and don't have enough account memory, API returns an error (no queueing runs).

Perhaps there is a simple solution by adding something like waitForAvailableMemory: true (or waitForAvailableMemoryIntervalSecs: 5) to API clients which would parse if the API error is about memory and then tried again. We could copy paste docs to users as a "local queue emulation" and it would also provide some documentation that this needs to be managed on the client.

Discussed on https://apify.slack.com/archives/CD0SF6KD4/p1766395205148049

[barjin]

Afaik there were discussions about adding run queue to the platform, no? I'd rather see this solved on the API / Platform level and keep the client as lean as possible. Implementing an automatic polling mechanism adds extra complexity, both implementation- and interface-wise (how should e.g. the run timeout option work with this?)