[8.11] Fix pipeline name in generated ingest processors (#169163)

## Summary

When creating an inference pipeline in Search, pipeline name references
in the metadata of the generated pipeline definition are incorrect; the
`ml.inference.` prefix is missing. This PR fixes that.

There are no functional side effects, since the pipeline name only
appears in metadata and failure messages generated during ingestion.

Affected fields in the pipeline definition:
- `inference.on_failure[].append.value.message`
- `inference.on_failure[].append.value.pipeline`
- `append.value.pipeline`

Without fix:

![274963743-743cac19-2543-4005-984b-a47468d3039a](https://github.com/elastic/kibana/assets/14224983/d19cef8d-672c-4eda-90cc-7d67e155ffe1)

With fix (definition / indexed document):
![Screenshot 2023-10-17 at 14 45
26](https://github.com/elastic/kibana/assets/14224983/e044482e-36ef-42ce-8383-07b31c98e46d)
![Screenshot 2023-10-17 at 14 47
54](https://github.com/elastic/kibana/assets/14224983/91204fa7-9148-4c47-a7b9-63688a3420ab)

Co-authored-by: Kibana Machine <42973632+kibanamachine@users.noreply.github.com>

x-pack/plugins/enterprise_search/public/applications/enterprise_search_content/components/search_index/pipelines/ml_inference/ml_inference_logic.ts

@@ -15,6 +15,7 @@ import {
   generateMlInferencePipelineBody,
   getMlInferencePrefixedFieldName,
   getMlModelTypesForModelConfig,
+  ML_INFERENCE_PREFIX,
   parseMlInferenceParametersFromPipeline,
 } from '../../../../../../../common/ml_inference_pipeline';
 import { Status } from '../../../../../../../common/types/api';
@@ -533,7 +534,7 @@ export const MLInferenceLogic = kea<
 
         return generateMlInferencePipelineBody({
           model,
-          pipelineName: configuration.pipelineName,
+          pipelineName: `${ML_INFERENCE_PREFIX}${configuration.pipelineName}`,
           fieldMappings: configuration.fieldMappings ?? [],
           inferenceConfig: configuration.inferenceConfig,
         });