redis: use pipeline when calling list_metric_with_measures_to_process

[gordon chung]

if there are a lot of buckets, we'll make a lot of calls to each bucket individually. this works but it'd be better if we can pass in a group of buckets so we can get them in a single connection using pipeline to minimise load.

this doesn't work for any of the other backends.

[gordon chung (chungg)]

i don't think this is relevant since we are dropping scheduler process