Deprecate `size` setting

[jordansissel]

The size setting is how many records are returned per scroll request, and it can easily be confused with "how many records to return, period" which is not what it does.

In practice, users should not care about this setting, and I find no reason why a user would want to set it.

The default (1000) is probably fine for every use case.

[ctindel]

@jordansissel do you think users should need to change any of the scroll settings? Or is the use case always "process every record returned by the query no matter how long it takes"? For example I always use 60000m in mine.

[jordansissel]

The scroll setting (like scroll: 1m) is for each request not the entire life of the scan. Basically you want to call it before the next expiration. So if you have default 1000 size and 1m you need to finish Each 1000 docs per call in less than 1 minute otherwise the handle in ES expires. 60m for 1000 records (per call) seems too high (that’s really slow events rate planning)

…

On Sat, Nov 18, 2017 at 8:15 AM Chad Tindel ***@***.***> wrote: @jordansissel <https://github.com/jordansissel> do you think users should need to change any of the scroll settings? Or is the use case always "process every record returned by the query no matter how long it takes"? For example I always use 60000m in mine. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#85 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAIC6l68X_TcQHzbhtfDUWWtWQY-VqFBks5s3wKdgaJpZM4Qi2m0> .

[SyedShahzaib790]

Hi,
Can someone tell me, whats the valid parameter to use in order to get specific number of records i.e. 10 records only. As it is mentioned above, one can confuse this size parameter. I am one of the victim :p
Can please specify which parameter/attribute to use to get specific number of records instead?