Grafana Loki Log Aggregation for Incident Investigations
summary
Loki is a log store that ties logs to Prometheus-style service discovery so logs and metrics share labels, letting investigations follow the exact instance that showed a spike in 99th percentile tail latency. It trades full-text inverted indexes for only index metadata, which claims much lower cost. It ingests pod stdout over gRPC streams, runs as Cortex-based multi-tenant microservices, and surfaces errors like 500s, lock timeouts and connection limits alongside dashboards.
AI-written summary · about
related talks
25mLoki: Prometheus-inspired logging for cloud native
31mLoki: Like Prometheus But for Logs
37mGrafana Loki: Like Prometheus, But for logs.
53mGetting started with Grafana Loki, a modern logs database
44mIntroduction to Grafana Loki
15mOperations report: Running Loki on Different On-Prem Environments
27mObservability with Prometheus and beyond
29m5 Years in, Loki Turns 3.0