Project: Log analyzer CLI

You're an SRE at esc bash. Every service on the fleet writes a plain text log, and when something breaks the first question is always the same: how many warnings and errors are in here, and how bad is it? Reading the file by eye does not scale. Build a small command line tool that turns a log into a level by level summary you can drop into a report or an alert.

Set up the project

Run this first. It seeds the input file for the task:

curl -fsSL https://raw.githubusercontent.com/Esc-Bash/project-init-scripts/main/python-for-devops/project-1/init.sh | bash

This creates /root/mission/app.log. Every line starts with a timestamp, then a log level, then a message, for example:

2026-07-24 10:00:07 INFO request handled ok

The level is always one of DEBUG, INFO, WARNING, ERROR, or CRITICAL.

What to build

Write /root/scripts/loganalyzer.py. It takes the path of a log file as a positional argument and the path of the summary to write as an -o/--output option:

loganalyzer.py <input-log> -o <output-json>

The script must:

1. Exit with code 2 if the log path argument is missing. argparse gives you this behaviour when the argument is required.
2. Exit with code 1, after printing an error to stderr, if the input log file does not exist.
3. Otherwise read the log, use a regular expression to pull the level out of each line, and count how many lines carry each level.

Write the result to the output path as JSON with this exact shape:

{
  "total": <number of log lines with a level>,
  "levels": {
    "DEBUG": <count>,
    "INFO": <count>,
    "WARNING": <count>,
    "ERROR": <count>,
    "CRITICAL": <count>
  }
}

Run your tool against the seeded log, writing the summary to /root/answers/summary.json:

loganalyzer.py /root/mission/app.log -o /root/answers/summary.json

Press Submit when done.