Website

Category

Next app

Zebrium

Using GPT-3 for plain language incident root cause from logs

About Zebrium

Zebrium is an innovative company that utilizes unsupervised machine learning to quickly and accurately identify the root cause of an incident.

Recently, their CTO wrote about a new feature that takes advantage of GPT-3 language model technology. This new feature uses GPT-3 to produce succinct, plain language root cause summaries from log data. Zebrium's unsupervised ML is already capable of finding the root cause and generates reports with 5-20 log events that identify the first event in the sequence (usually the root cause), the worst symptom, other connected events, and correlated metric anomalies.

By integrating GPT-3, Zebrium can now take the next step and distill the root cause reports into concise natural-language summaries. This is done by searching the internet for similar incidents and extracting short "English" descriptions for users to easily digest.

You can read the full blog post here: https://www.zebrium.com/blog/real-world-examples-of-gpt-3-plain-language-root-cause-summaries-zebrium

Zebrium screenshots

Zebrium - screen 1
Zebrium - screen 2
Zebrium - screen 3

Read in Ukrainian or Ru