Problem
Operator appends all logs to a single file (~/.operator/logs/operator.log) with no rotation. On a moderately active install (4 agents, 6 jobs), this grows ~1.4MB/day. Over time the log becomes unwieldy and wastes disk.
Proposal
Add built-in log rotation to the Operator runtime with a config option:
logging:
max_size: "10MB" # rotate when file exceeds this size
max_files: 7 # keep N rotated files
compress: true # gzip old files
Sensible defaults so it works out of the box without config. Python's logging.handlers.RotatingFileHandler would handle this natively.
Current workaround
External logrotate config or manual truncation in a cron/backup script.
Problem
Operator appends all logs to a single file (
~/.operator/logs/operator.log) with no rotation. On a moderately active install (4 agents, 6 jobs), this grows ~1.4MB/day. Over time the log becomes unwieldy and wastes disk.Proposal
Add built-in log rotation to the Operator runtime with a config option:
Sensible defaults so it works out of the box without config. Python's
logging.handlers.RotatingFileHandlerwould handle this natively.Current workaround
External logrotate config or manual truncation in a cron/backup script.