| name | log-parser |
| description | Parse structured JSON-lines log files, filter by time range and severity, output as CSV. |
| version | 1.0.0 |
Log Parser
Skill Summary
Parses structured log files in JSON-lines (.jsonl) format. Each line is a JSON
object with at least timestamp, level, and message fields. The skill filters
events by time range and/or severity level and writes the matching rows as a CSV
file, making downstream tabular analysis easy.
Inputs
| Parameter | Flag | Required | Description |
|---|
| input | -i | Yes | Path to the .jsonl log file |
| output | -o | Yes | Path to the output CSV file |
| start-time | --start | No | Start of time window (ISO-8601, e.g. 09:00) |
| end-time | --end | No | End of time window (ISO-8601, e.g. 17:00) |
| levels | --levels | No | Comma-separated severity filter (e.g. ERROR,WARN) |
Processing Steps
- Read the input file line by line; parse each line as JSON.
- Normalise
timestamp to datetime; skip malformed lines.
- Apply time-range filter (if
--start / --end provided).
- Apply severity filter (if
--levels provided; case-insensitive).
- Write surviving records as CSV with columns:
timestamp, level, service, endpoint, response_time_ms, status_code, message
Output
A CSV file containing the filtered log events. Columns are derived from the JSON
keys present in the log entries.
Implementation
python3 ./skills/log-parser/log_parser.py -i <input.jsonl> -o <output.csv> \
[--start HH:MM] [--end HH:MM] [--levels LEVEL1,LEVEL2]
Dependencies
- Python 3.10+ (standard library only)
