New Keyword App

One of the most common requests I get from new customers is that they want to centrally collect all their machine generated time series data and search for a keyword like error or RuntimeException. Obviously Splunk can do this. Then, the next set of questions concern things like give me the top hosts or applications producing this keyword, show me a baseline of last week vs this week for this keyword, show me a slope line on the trend for this or any keyword(s), find outliers that go beyond the average occurrences for the keyword and then try to predict what may happen in the future.

To answer these questions and then some, I’ve created an app template that you can download from Splunkbase and simply install the Keyword App. The app has no inputs as it uses your own indexed data. You can modify the default indexes for the role that uses the app in Manager to include indexes beyond main. Usually, what I do next in these blog entries is describe how to use the app. However, in this case, a picture is worth a thousand words. Rather than continue to entertain you with my prose, I’ll simply embed the the self explanatory dashboards here.

Top Host, Sourcetype, Source

Top Host, Sourcetype, and Source for Keyword

Comparing Present Count vs. Past Count

Slope for Trend

Find Outliers

Predict Future Counts (Recession Predict is Jack Coate’s idea)

There’s a Donut Chart (Ron Naken’s Module) that allows you to split your list of keywords by top 10 host, source, or sourcetypes and see the distribution of events.

Split Keyword(s) Distribution by Top 10 Hosts, Source, or Sourcetype

Here are rare events based on your filter:

Rare

Finally, here are rare events based on punctuation.

Punct

Conclusion

Hopefully, this template will satisfy all your needs out of the box. If you need to modify the searches, make a copy of the app’s default/data/ui/views/<dashboard>.xml file and put it into the app’s local/data/ui/views/<dashboard>.xml file and edit your local copy. I’m hoping this app will save you some time from configuring these common tasks.

Related Articles

Deep Learning in Security: Text-based Phishing Email Detection with BERT Model
Security
4 Minute Read

Deep Learning in Security: Text-based Phishing Email Detection with BERT Model

We introduced a large language model (LLM)-based phishing email detector integrated into the Splunk DSDL app. We provide details on model training and evaluation, comparisons to other machine learning and deep learning algorithms as well as deployment approaches to Splunk in this blog.
Cybersecurity Awareness Month Spotlight: Insights from the Cisco Talos & SURGe Teams
Security
3 Minute Read

Cybersecurity Awareness Month Spotlight: Insights from the Cisco Talos & SURGe Teams

The Cisco Talos and SURGe by Splunk teams gathered for a special episode of Talos Takes filled with engaging cybersecurity discussions and candid opinions.
Monitoring Pulse Connect Secure With Splunk (CISA Emergency Directive 21-03)
Security
11 Minute Read

Monitoring Pulse Connect Secure With Splunk (CISA Emergency Directive 21-03)

Our Splunk security experts share a closer look at the Pulse Connect Secure attack, including a breakdown of what happened, how to detect it, and MITRE ATT&CK mappings.