documentation on checkpointInterval for connectedComponents is insufficient

[fritshermans]

The api documentation on connectedComponents is not clear about what the checkpointInterval argument is affecting:
"checkpointInterval – checkpoint interval in terms of number of iterations (default: 2)"
Does the value of this argument have impact on performance? On actual results?

[uazafar]

Based on the usage of the variable checkpointInterval elsewhere in Spark, this implies the frequency at which we cache the results but it requires a checkpoint directory to be set in the SparkContext. So there will be a performance hit I think.

I'm assuming it works in the same way as:
(https://spark.apache.org/docs/latest/api/python/reference/api/pyspark.ml.recommendation.ALS.html?highlight=checkpoint#pyspark.ml.recommendation.ALS.checkpointInterval)