Splunk Search

How do you make output from xyseries generate the same _time-based X-axis labels as timechart?

The following search:

sourcetype=my_log_type | timechart count by conn_type

generates the chart I want, with one exception: instead of the original conn_type (connection type) values, I want the chart to display more readable, descriptive values.

I want to replace the values after the timechart command. For reasons why, see my comment on a different question.

To rename the series, I append the following commands to the original search:

| untable _time conn_type value | lookup connection_types.csv conn_type output description | xyseries _time description value

This has the desired effect of renaming the series, but the resulting chart lacks the intelligently formatted X-axis values generated by timechart.

How do I reproduce the intelligent X-axis values generated by timechart?

I do not want to hardcode span or bin values: I use this search in a dashboard whose time range might span years or fractions of a second. I want to take advantage of the way timechart chooses a default span based on the time range.

The following screenshots show the original "intelligent" X-axis labels generated by timechart, and the X-axis labels after xyseries:

Splunk screenshots

1 Solution

Motivator

What about doing a stats and then a timechart?

sourcetype=my_log_type | bucket _time span=1s | stats count by _time, conn_type | lookup connection_types.csv conn_type output description | timechart sum(count) AS count by description

View solution in original post

Contributor

I personally can not use timechart. Is there a way to do what timechart is doing with xyseries ?

EDIT: I found a solution:

I perform the xyseries, then the "untable", and then the timechart... But it really not efficient and optimized. That would be great to know exactly what the timechart command is doing...

0 Karma

Motivator

What about doing a stats and then a timechart?

sourcetype=my_log_type | bucket _time span=1s | stats count by _time, conn_type | lookup connection_types.csv conn_type output description | timechart sum(count) AS count by description

View solution in original post

Thank you, @Jeremiah! That works for me.

I've removed the span=1soption after reading the docs: bucket (and bin) seem to share the same default spanning behavior as timechart. I've also replaced the bucket command name with bin, because - tell me if I'm wrong - the bin command seems to be the "primary" command (for which bucket is an alias): the Splunk docs topic for bucket refers the reader to the bin topic.

I'd like to convert your comment into an answer so that I can accept it, but I can't see how to do that. I'm guessing I lack the authority - or karma points - for that option to appear in my user interface. Could you (or anyone reading this) please do that for me, or point me to where I can do that myself?

So, pushing timechart to the end of the search solves my problem. I'm still curious, though: timechart seems to be "doing stuff under the covers" (perhaps: generating "internal use only fields" that Splunk "hides" from users?) that I do not (yet?) have the wit to see.

0 Karma

Community Manager
Community Manager

Looks like it is now an answer 🙂 accept away!

0 Karma