Labor XI · Chapter 11

Powering Up Your Skills

Use AI assistants to write complex time-series queries in seconds, and learn how to prompt them precisely enough to trust the result.

View or suggest edits on GitHub

Building visualizations using ChatGPT#

Building time series visualizations can be challenging but you can use ChatGPT to speed things up. Say that you wanted to identify NTP traffic from devices on your network.

  1. Go to the ChatGPT web interface.
  2. Type:
I want to visualize NTP traffic from devices on my network using Timelion on Kibana. My index is logstash*.

ChatGPT is more familiar with Kibana’s Timelion than Opensearch’s Timeline, but they are the same thing. 3. Its response is:

.es(
  index=logstash*, 
  q='event_type:flow AND (dest_port:123 OR src_port:123)', 
  metric='count'
).label('NTP Traffic Count')

Just make sure to be as specific as possible when asking ChatGPT for help.

Here’s the image of the visualization: NTP traffic count