Version 0.5 is our biggest Streamline release so far, focused on reliability and observability. The core queueing model is unchanged, but the operator workflow is much smoother for real production traffic.
Multiple consumers can now share work predictably with explicit group offsets. This enables horizontal scaling without custom coordination logic in every service and improves throughput consistency during bursty workloads.
Failed events can be routed automatically after configurable retry windows. Operators get clear visibility into poison messages while the primary stream stays healthy. Replay tools are included for controlled reprocessing.
The new dashboard surfaces lag, retries, and consumer health in one view. Instead of stitching metrics from several tools, teams can inspect stream behavior quickly and act before small issues become incidents.