Datehistogram + statistical facet: variable time series intervals & aggregates in ES


(Michael Sick) #1

I have a stream of timeseries data that streams to ES once / device /
second. I would like to apply a facet (or aggregate) that buckets based on
time but produces the outputs of the stats facet on a value during that
time.

So something like 5 minute buckets from start_date to end_date
giving count, total, sum of squares, mean (average), minimum, maximum,
variance, and standard deviation values for those buckets.

Can this be done in the 0.9.x branch? If so, guidance on how and tradeoffs
will be helpful. How about 1.x and the new aggregates? Client is not going
live for a few months so the 1.x branch could be reasonable.

Thanks,

--Mike

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAP8axnDbqOfaexQDEDSe-R%3DQ_Kx-ccsvqcaE7Y13NVnuDxJ3QQ%40mail.gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.


(Steve Mayzak) #2

Michael,

Thats exactly what the new Aggregations feature is for. Have you seen our
recent blog post for the 1.0 beta 2 release? Aggregations is included in
that release.

I would download the beta and try your example there and let us know how it
goes.

Here are the aggregations docs:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/master/search-aggregations.html

On Monday, December 9, 2013 1:38:12 PM UTC-8, Michael Sick wrote:

I have a stream of timeseries data that streams to ES once / device /
second. I would like to apply a facet (or aggregate) that buckets based on
time but produces the outputs of the stats facet on a value during that
time.

So something like 5 minute buckets from start_date to end_date
giving count, total, sum of squares, mean (average), minimum, maximum,
variance, and standard deviation values for those buckets.

Can this be done in the 0.9.x branch? If so, guidance on how and tradeoffs
will be helpful. How about 1.x and the new aggregates? Client is not going
live for a few months so the 1.x branch could be reasonable.

Thanks,

--Mike

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/65fa4aa3-9579-4f83-b5fd-01a94b06cf5e%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.


(system) #3