[BUG] Error in Delete ML Model Runner

[vpehkone]

Describe the bug
The deleting ML model can fail, if the undeploying model has not yet completed.

To Reproduce
Run DeleteMlModel runner, when there is already registered and deployed ML model.
Use workload to run it: Run the workload treccovid_semantic_search (this workload is not yet merged to main repo) two times. In the second run, the task delete-ml-model can fail.

Expected behavior
If there is already register and deployed model, it will be undeployed and deleted.

More Context (please complete the following information):

• Workload: https://github.com/vpehkone/opensearch-benchmark-workloads/tree/neural_search/treccovid_semantic_search