Splunk Search

How to extract a field with regex

maria_n
Explorer

Hi Everyone

Sample logs:

{"kubernetes":{"container_name":"sign-template-services","namespace_name":"merch-ps-signs-stress-1","pod_name":"sign-template-services-14-chfbn"},"message":"::ffff:100.65.19.1 - - [05-Mar-2020 09:58:48 CST] \"GET /health HTTP/1.1\" 200 30 - **7.807** ms\n","hostname":"ocp-usc1-lle-b-app-f-g3q9.c.kohls-openshift-lle.internal","@timestamp":"2020-03-05T15:58:48.231999+00:00","cluster_name":"ocp.gcpusc1-b.lle.xpaas"}

{"kubernetes":{"container_name":"sign-template-services","namespace_name":"merch-ps-signs-ci","pod_name":"sign-template-services-39-gb69d"},"message":"::ffff:100.109.92.1 - - [05-Mar-2020 09:57:31 CST] \"GET /health HTTP/1.1\" 200 30 - **33.245** ms\n","hostname":"ocp-usc1-lle-c-app-f-7ml9.c.kohls-openshift-lle.internal","@timestamp":"2020-03-05T15:57:31.808739+00:00","cluster_name":"ocp.gcpusc1-c.lle.xpaas"}

We need to extract a field called "Response_Time" which is highlighted in these logs. The data is available in the field "message".
I have tried the below regex but it does not seem to work.

index=kohls_prod_infrastructure_openshift_raw kubernetes.container_name=sign-template-services 
| rex field=MESSAGE "\d{3} d{2} - (?\d+) ms\""

Please help!

Thanks.

0 Karma
1 Solution

manjunathmeti
Champion

hi @maria_n,

Try this:

index=kohls_prod_infrastructure_openshift_raw kubernetes.container_name=sign-template-services 
| rex field=message "-\s(?<Response_Time>[\d\.]+)"

Sample query:

| makeresults 
| eval message = "::ffff:100.65.19.1 - - [05-Mar-2020 09:58:48 CST] \"GET /health HTTP/1.1\" 200 30 - 7.807 ms\n" 
| rex field=message "-\s(?<Response_Time>[\d\.]+)"

View solution in original post

manjunathmeti
Champion

hi @maria_n,

Try this:

index=kohls_prod_infrastructure_openshift_raw kubernetes.container_name=sign-template-services 
| rex field=message "-\s(?<Response_Time>[\d\.]+)"

Sample query:

| makeresults 
| eval message = "::ffff:100.65.19.1 - - [05-Mar-2020 09:58:48 CST] \"GET /health HTTP/1.1\" 200 30 - 7.807 ms\n" 
| rex field=message "-\s(?<Response_Time>[\d\.]+)"

darrenfuller
Contributor

you didn't give the field a name and you didn't capture the period. Try something like:

rex field=message"\d{3}\s\d{2}\s\-\s(?<Response_Time>[\d\.]+)\sms"

./DF

gcusello
SplunkTrust
SplunkTrust

Hi @maria_n,
only one question: the field is "message" or "MESSAGE"? check the correct case.
try something like this:

index=kohls_prod_infrastructure_openshift_raw kubernetes.container_name=sign-template-services
| rex field=MESSAGE "\s+(?<Response_Time>[^ ]+)\sms"

you can test the regex at https://regex101.com/r/82WdWC/1

Ciao.
Giuseppe

Get Updates on the Splunk Community!

Hunt Smarter, Not Harder: Discover New SPL “Recipes” in Our Threat Hunting Webinar

Are you ready to take your threat hunting skills to the next level? As Splunk community members, you know the ...

Splunk ITSI & Correlated Network Visibility

  Now On Demand   Take Your Network Visibility to the Next Level In today’s complex IT environments, ...

Community Content Calendar, August edition

In the dynamic world of cybersecurity, staying ahead means constantly solving new puzzles and optimizing your ...