TestMultipleVMs_Isolated is still flaky

https://buildkite.com/firecracker-microvm/firecracker-containerd/builds/2535#8267eaab-d952-4276-a956-c304790b50b0

```
=== RUN TestMultipleVMs_Isolated
service_integ_test.go:263: TestMultipleVMs_Isolated: will run 100 vm's
service_integ_test.go:450:
Error Trace: service_integ_test.go:450
Error: Received unexpected error:
unexpected errors from VM 76: failed to create a VM while there are 0 Firecracker processes: rpc error: code = Unknown desc = failed to create VM: failed to dial the VM over vsock: non-temporary vsock dial failure: failed to dial "rootfs/firecracker.vsock" within 100ms: dial unix rootfs/firecracker.vsock: connect: resource temporarily unavailable
Test: TestMultipleVMs_Isolated
--- FAIL: TestMultipleVMs_Isolated (132.93s)
```

The whole error message doesn't sound right.

> **non-temporary** vsock dial failure: failed to dial "rootfs/firecracker.vsock" within 100ms: dial unix rootfs/firecracker.vsock: connect: resource **temporarily** unavailable

Start with TestMultipleVMs_Isolated in service_integ_test.go, especially the failure reported at lines 263 and 450, and review the vsock dialing path named in the failure. Reproduce the flaky run if possible, then verify that the error classification and wording correctly describe the reported temporary resource-unavailable condition.

Written by the indexing model from the issue text.