5 Years in, Loki Turns 3.0
summary
Loki 3.0 ships with native OTel support, a bloom-filter based “sketch” index that automatically rejects useless chunks, and a new Explore Logs UI so developers can find issues without learning LogQL. Internal runs show a cell ingesting ~43 TB/day with only ~200 MB of index and bloom artifacts at roughly 2–3% of log size, letting queries discard ~75–90% of data before scanning. Releases also clean configs, improve Helm installs, and remove legacy storage.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:46Reflecting on Loki - how did we get here?
- 3:23Minimal indexing
- 6:40Loki 3.0: Schemaless at scale
- 8:51Schemaless query acceleration
- 12:25Bloom filters
- 17:38Loki and OTel
- 19:08Loki cleanup
- 19:42Loki and Helm
- 22:31Explore Logs
- 23:00Demo
related talks
31mLoki: Like Prometheus But for Logs
37mGrafana Loki: Like Prometheus, But for logs.
44mIntroduction to Grafana Loki
05mGrafana Loki Log Aggregation for Incident Investigations
53mGetting started with Grafana Loki, a modern logs database
25mLoki: Prometheus-inspired logging for cloud native
28mFrom Fragmented Logs to Unified Insights: Scaling O11y
35mFrom New Cluster to Insight. Deploying Monitoring and Logging to Kubernetes