[DOCS] Skips testing for forecast API

[lcawl]

This PR skips code snippet testing for the machine learning forecast API, since (even though it works from the console with basic farequote data), the gradle checks fail as follows:

1> [2018-06-18T20:45:24,050][INFO ][o.e.s.XDocsClientYamlTestSuiteIT] Stash dump on test failure [{
1> "stash" : {
1> "body" : {
1> "error" : {
1> "root_cause" : [
1> {
1> "type" : "status_exception",
1> "reason" : "Cannot run forecast: Forecast cannot be executed as job requires data to have been processed and modeled",
1> "stack_trace" : "ElasticsearchStatusException[Cannot run forecast: Forecast cannot be executed as job requires data to have been processed and modeled]

I suspect the CI is too quick and the setup data has not been modelled yet.

[elasticmachine]

Pinging @elastic/ml-core

[hendrikmuhs]

I run into a similar issue recently, in automation things are simply to quick. I think if you flush the job prior forecasting it should start working, something like:

  - do:
      xpack.ml.flush_job:
        job_id: "total-requests"

[lcawl]

@hendrikmuhs Unfortunately that doesn't seem to do the trick. If you don't have any other suggestions, I'll just disable these tests.

[hendrikmuhs]

@lcawl

I checked again: The problem is the datafeed. Unfortunately I could not find a way to wait for the completion of the datafeed. I missed that flush is not working for this case, sorry, my mistake.

There is 1 more possibility, if you use xpack.ml.post_data instead of a data feed, which means directly uploading data into a job by-passing indexing and data-feeding and again wait for the flush.

But: I guess this would be a big change, likely affecting other parts or you need a separate jobs to test forecasting.

I do not know if it's worth the effort, please judge yourself, I am fine with disabling the test if you prefer.

[lcawl]

@hendrikmuhs I unfortunately encountered the same problem with posting data, so I will leave this test disabled. Thanks for the suggestions!