Missing idleTimeout key in daskcluster_autoshutdown

**Describe the issue**:

My KubeClusters sometimes do not get shut down properly on kubernetes when they're done with their work. Kubernetes logs state that there's an exception in a kopf finalizer which is retried indefinitely, apparently due to the spec dict given to `daskcluster_autoshutdown`:

```
Timer 'daskcluster_autoshutdown' failed with an exception. Will retry.
Traceback (most recent call last):
File "/usr/local/lib/python3.10/site-packages/kopf/_core/actions/execution.py", line 276, in execute_handler_once
result = await invoke_handler(
File "/usr/local/lib/python3.10/site-packages/kopf/_core/actions/execution.py", line 371, in invoke_handler
result = await invocation.invoke(
File "/usr/local/lib/python3.10/site-packages/kopf/_core/actions/invocation.py", line 116, in invoke
result = await fn(**kwargs) # type: ignore
File "/usr/local/lib/python3.10/site-packages/dask_kubernetes/operator/controller/controller.py", line 852, in daskcluster_autoshutdown
if spec["idleTimeout"]:
File "/usr/local/lib/python3.10/site-packages/kopf/_cogs/structs/dicts.py", line 297, in __getitem__
return resolve(self._src, self._path + (item,))
File "/usr/local/lib/python3.10/site-packages/kopf/_cogs/structs/dicts.py", line 121, in resolve
result = result[key]
KeyError: 'idleTimeout'

```

When I remove these lines from the DaskCluster resource YAML in kubernetes, the problem is gone

```yaml
finalizers:
- kopf.zalando.org/KopfFinalizerMarker
```

Is it correct that `daskcluster_autoshutdown` as below receives `spec` as a specification dict, e.g. from `make_cluster_spec(..., idle_timeout=5)`? I tried expicitly adding the `idle_timeout`, but the problem persists

```python
@kopf.timer("daskcluster.kubernetes.dask.org", interval=5.0)
async def daskcluster_autoshutdown(spec, name, namespace, logger, **kwargs):
if spec["idleTimeout"]:
try:
idle_since = await check_scheduler_idle(
scheduler_service_name=f"{name}-scheduler",
namespace=namespace,
logger=logger,
)
except Exception:
logger.warn("Unable to connect to scheduler, skipping autoshutdown check.")
return
if idle_since and time.time() > idle_since + spec["idleTimeout"]:
cluster = await DaskCluster.get(name, namespace=namespace)
await cluster.delete()
```

Not sure if this is a proper bug, or an issue with kopf, or anything is misconfigured on my end. Appreciate any help.
I'd also be fine with just removing the timer/finalizer if that's possible.

**Anything else we need to know?**:

**Environment**:

- Dask version: 2024.4.1
- Dask operator version: 2024.4.0
- Python version: 3.10.12
- kopf python version: 1.37.1
- Operating System: Linux
- Install method (conda, pip, source): pip

Start in dask_kubernetes/operator/controller/controller.py at daskcluster_autoshutdown and trace how the spec is built by make_cluster_spec(..., idle_timeout=5). Reproduce the KeyError with the reported Dask and operator versions, including the finalizer in the resource YAML. Done means autoshutdown handles the missing idleTimeout case without an indefinitely retried finalizer failure.

Written by the indexing model from the issue text.