Loki Pairs With Kafka and Smarter Storage to Help Calm Cost and Scaling Challenges
Techstrong IT, Monday, May 4th, 2026
Grafana's Loki now uses Kafka and columnar storage to reduce logging costs by 30% and improve query performance 10x.
Grafana Labs unveiled significant improvements to its open-source Loki logging project at GrafanaCon, integrating Apache Kafka as a new orchestration mechanism for distributed systems to achieve better scalability and a 30% cost reduction. The update introduces columnar storage format (data objects) that replaces the previous row-based chunks format, enabling selective reading of specific metadata without downloading entire log files.
New search optimization using Bloom filters and an improved query engine allow queries to scan selectively over 20 times less data while delivering 10 times faster performance on aggregated queries. These architectural changes address the explosion of log data volumes and redundancy challenges that have plagued large-scale distributed logging operations.