improve handling of graceful abort when runs are about to be started

[halvko]

If the run is gracefully aborted during a call to client.actor(...).start(...), the orchestrator crashes with

file:///usr/src/app/node_modules/apify-orchestrator/dist/clients/actor-client.js:103
            throw new Error(`Error starting Run: ${runName} (${this.id}).`);
                  ^

Error: Error starting Run: run-name (scraper/name).
    at ExtActorClient.enqueueAndWaitForStart (file:///usr/src/app/node_modules/apify-orchestrator/dist/clients/actor-client.js:103:19)
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)

after investigating I found no extra runs started or anything, and everything worked correctly after resurrection. I think how scary the error message looks is unfortunate, but I'm also uncertain if it can behave wrongly in cases where multiple start calls are done concurrently (e.g.

const [run1, run2] = Promise.all([
  actor.start('name1', ...),
  actor.start('name2', ...),
]);

or equivalent). If special care should be taken around actor.start (e.g. using Promise.allSettled or wrapping calls in try blocks) to ensure correct behavior I think it should at least be documented on the method 😌 .