InferenceLostConnection error for detect step with more than 5000 data points using Multivariate Anomaly Detection API in Python

> Please provide us with the following information:
> ---------------------------------------------------------------

### This issue is for a: (mark with an `x`)
```
- [x] bug report -> please search issues before submitting
- [ ] feature request
- [ ] documentation issue or request
- [ ] regression (a behavior that used to work and stopped in a new release)
```

### Minimal steps to reproduce
When I submit more than 5,000 data points for the multivariate anomaly detector `detect` API call [https://{endpoint}/multivariate/models/{model_id}/detect](https://%7Bendpoint%7D/multivariate/models/%7bmodel_id%7d/detect)), I consistently get the following error. According to the documentation, the maximum number of data points per inference call is 20,000. There is no error when I use less than 5,000 data points for the API call.

Supplying at most 5,000 data points is much more time-consuming because any single detect API call takes around the same time to complete (regardless of the number of data points) in my experience.

### Any log messages given by the failure
> Error code: InferenceLostConnection. Message: Lost connection with the worker that processed the inference work. This may be caused by service upgrade. Re-trigger inference should help resolve this.

### Expected/desired behavior
> No error when I supply at most 20,000 data points for a single detect API call.

### Mention any other details that might be useful