Log analysis using ml



I am trying to build a model which will auto-triage the issue,if it finds an existing problem inside production java logs(I mean based on most recent exception found inside logs),so It will send out a note that some exception found which may be environment issue,infra issue or actual bug.

For doing that,I had created a table in db in which I had entered few types of exceptions and corresponding issue type(infra/bug/env). Planning to read the logs from bottom and want to analyse root cause automatically.

Need inputs on this ,how to use ml here and auto-train the model if nay new issue is hit. Thanks in advance