Timelion Time-Series data based on field value


I am attempting to view a time series plot of data using timelion.

I have roughly 2500 samples of a signal strength value (RSSI) taken over a over a 5 day period - for each frequency (roughly 560 different frequencies in my index). The sample time is imported into Kibana as "date" field and is being used as the current default timefield.

I am struggling to just view the RSSI values for a particular frequency versus time. I do not want to view a sum or average - simply examine the signal strength as time progresses and look for correlations.

I cannot seem to get this to work. Any ideas here?

Individual samples are roughly 3 min apart from one another.

Hi Brandon, have you tried something like this:

.es(index=your-index-pattern-*, timefield='date'', metric='avg:{name of field for frequency A}'),
.es(index=your-index-pattern-*, timefield='date'', metric='avg:{name of field for frequency B}'),
.es(index=your-index-pattern-*, timefield='date'', metric='avg:{name of field for frequency C}')

I know you stated that you don't want to view an average, but that's how Elasticsearch works. It extracts data by forming bucket aggregations (in this case across time), and then calculating values based on the documents that fall into that bucket.

We can refine the data you're seeing in Timelion by normalizing your time series and by scaling the interval down to a faster rate (e.g. 1 second or 1 minute), so that the visualized bucket size is smaller. Take a look this this awesome Timelion tutorial for more info.

Alternatively, you could try out our experimental Time Series Visual Builder (available under Visualize). We have a couple tutorials available on our blog (one and two).

Please let me know if this helps!


Hey CJ,

Thank you for your quick and detailed response. I was able to correct my issue. The timefield='date' argument was hung up. Restarting the Kibana service allowed me to correctly view the data as described.

Further reading informed me that I should be using the "max" metric to view the sampled data versus time; avg is also useful. I will look at the other tutorials and blog suggestions you provided.

Thanks much!

Oh, great! I'm glad that solved the problem -- thanks for sharing it. :slight_smile:

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.