Configure retries for a Trigger.dev v4 task, abort retries deliberately, and customize retry behavior with the catchError hook

Documented steps

1. Set per-task retry config: `retry: { maxAttempts: 10, factor: 1.8, minTimeoutInMs: 500, maxTimeoutInMs: 30_000, randomize: false }` on the `task()` definition (https://trigger.dev/docs/errors-retrying).
2. Set project-wide retry defaults in trigger.config.ts; task-level `retry` overrides them.
3. To stop retrying from inside a run, throw `AbortTaskRunError` from `@trigger.dev/sdk`: `throw new AbortTaskRunError("invalid input")`.
4. Add a `catchError` hook to intercept failures: `catchError: async ({ payload, error, ctx, retryAt }) => { if (permanent) return { skipRetrying: true }; if (later) return { retryAt: someDate }; }` — return `undefined` to fall back to default behavior.
5. Note the v4 rename: `handleError` is now `catchError`, and lifecycle hooks take a single object argument (https://trigger.dev/docs/upgrade-to-v4).
6. For OOM failures specifically, configure `retry: { outOfMemory: { machine: "large-1x" } }` so the retry lands on a bigger machine (https://trigger.dev/docs/machines).
7. Verify attempt behavior on the run detail page in the dashboard before shipping.