instability in e2e tests

There are definitely some instability but I don't believe it is related to these changes or https://github.com/containerd/runwasi/pull/631:

2 out of three pods started in the e2e (not oci artifact related):

```
default wasi-demo-6c5dcf6ddb-47qjh 2/2 Running 0 [8](https://github.com/containerd/runwasi/actions/runs/10326209465/job/28589545828?pr=661#step:8:9)2s
default wasi-demo-6c5dcf6ddb-6qh8s 2/2 Running 0 82s
default wasi-demo-6c5dcf6ddb-b6trr 0/2 ContainerCreating 0 82s
```

looking into the logs the pod with wasm started: `default_wasi-demo-6c5dcf6ddb-b6trr_c486fffe-6491-4731-af1c-a44df3603288` but the other one seems to have errors:

```
Aug 09 22:07:29 containerd-wasm-control-plane containerd[210]: time="2024-08-09T22:07:29.871600750Z" level=error msg="collecting metrics for e1788b4f7ae9ee221c7a5196d62b05e5c50572009d750438c240215ca43a21d1" error="task is not running: invalid argument"
Aug 09 22:07:03 containerd-wasm-control-plane containerd[210]: time="2024-08-09T22:07:03.041875285Z" level=error msg="loading cgroup for 377" error="cgroups: cannot find cgroup mount destination"
```

_Originally posted by @jsturtevant in https://github.com/containerd/runwasi/issues/661#issuecomment-2278843950_