Error Log Analysis

This example uses TigerGraph database to represent log messages within their contexts and finds similar trouble Logs messages for a given query message.

This example selects a random vertex in the graph from a query and returns the top matching Logs based on cosine similarity. Solutions from the matching logs can help resolve current issues.

Instead of finding similarity with the direct one-hot word representation, we used GloVe Word Embeddings, which maps words into more meaningful space.

In General, finding Cosine Similarity on large dataset will take a huge amount of time on CPU. With the Xilinx Cosine Similarity Acceleration, it can speed up the process by multiple orders.

Run Jupyter Notebook

  • Install the Log Similarity Demo Plugin

(fpga)$ cd log_similarity
(fpga)$ su - tigergraph
Password:
(fpga)$ bin/install_udf.sh

  • Run the command below to start Jupyter Notebook

(fpga)$ cd jupyter-demo
(fpga)$ jupyter notebook log_similarity_TG_demo.ipynb

  • Follow the step-by-step instructions in the notebook once it is loaded in your browser.

This use case can be downloaded from Error Log Analysis Jupyter notebook on GitHub.