Errors in production do not arrive evenly spread; they arrive in clumps when something upstream falls over. The burst option clusters eight consecutive failures into the timeline, which is what an alerting rule actually has to detect.
What you can control
- Six formats covering web servers, structured logging and syslog.
- The burst option groups errors together instead of scattering them, so alerting thresholds get a realistic test.
- Trace and span IDs for testing distributed-tracing correlation.
- Failed requests carry realistic latency — hundreds of milliseconds to tens of seconds, not the same as successes.
Testing ideas
- Parsers. Generate combined, common, JSON, logfmt and syslog lines to test log shippers and grok or regex parsers.
- Bursts. Cluster errors into a burst to test alerting thresholds — real incidents arrive in clumps.
- Tracing. Include trace and span IDs to test correlation in your observability stack.
- Safe IPs. Client IPs come from the RFC 5737 documentation range, so sample logs can be shared freely.
What this is not
Timestamps are spread evenly across the window rather than following real traffic patterns. A log covering a full day will not show the daily peak your capacity planning depends on.