fix nil referencing issue in cadvisor during metric decoration

[movence]

Description:
cadvisor is throwing an error while decorating metrics with k8s stores. The issue caused the agent pods to go into CrashLoop by throwing error in the agent logs:

goroutine 1076 [running]:
github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver/internal/cadvisor.(*Cadvisor).GetMetrics(0xc001a80180)
	github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver@v0.89.0/internal/cadvisor/cadvisor_linux.go:231 +0x461
github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver.(*awsContainerInsightReceiver).collectData(0xc000b23d40, {0x4c3b158, 0xc000485ef0})
	github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver@v0.89.0/receiver.go:351 +0x83
github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver.(*awsContainerInsightReceiver).Start.func1()
	github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver@v0.89.0/receiver.go:172 +0xe5
created by github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver.(*awsContainerInsightReceiver).Start in goroutine 1
	github.com/open-telemetry/opentelemetry-collector-contrib/receiver/awscontainerinsightreceiver@v0.89.0/receiver.go:158 +0x113f

Testing:
Tested on a cluster that was exhibiting the issue.

Documentation:

[jefchien]

Why are we including nils in the result slice in the first place?

[movence]

This happens when pod information is not found in store (PodStore or ServiceStore) cache while decorating metrics with k8s metadata or service. It is unknown why corresponding pod information is missing for a metric, but this adds a safe guard around this edge case.