Loki is a horizontally-scalable, multi-tenant log aggregation system built in Go and inspired by Prometheus. Rather than performing full-text indexing on log contents, Loki stores compressed, unstructured logs while indexing only metadata labels, making it significantly cheaper to operate and simpler to maintain than traditional log aggregation systems. The system is designed to work seamlessly with Prometheus by using the same label-based approach for both metrics and logs, allowing users to switch between the two using identical label sets. This label-centric architecture makes Loki particularly well-suited for Kubernetes environments, where Pod labels and metadata are automatically scraped and indexed.
The Loki stack consists of three components: Alloy, which is the agent responsible for gathering logs and sending them to Loki; Loki itself, which stores logs and processes queries; and Grafana, which provides the interface for querying and displaying logs. Alloy replaced Promtail as the primary log collection agent, with Promtail now considered feature-complete while future development focuses on Alloy's expanded capabilities. Loki has native support in Grafana starting from version 6.0, enabling integrated observability workflows.
The repository shows active development and community engagement. GitGenius tracking data reveals that since July 2026, the project has grown from 28495 to 28496 stargazers. The issue and pull request response latency shows a median of 126.4 hours across 2185 tracked items, though the mean of 5953.6 hours indicates some longer-tail response times. The most frequently labeled issues are type/bug with 504 occurrences, area/helm with 313, and type/feature with 295, reflecting ongoing maintenance and enhancement efforts. JStickler emerges as the most active contributor with 3369 tracked events, followed by chaudum with 336 events and trevorwhitney with 147 events.
The repository's scope extends across multiple operational domains. Documentation covers API usage for ingesting logs, label configuration, operational procedures, Docker driver integration for direct container log shipping, LogCLI for command-line querying, Loki Canary for monitoring log delivery, and troubleshooting guidance. A significant operational note indicates that as of March 16, 2026, the Grafana Loki Helm chart is being forked to a new repository at grafana-community/helm-charts, with the original repository maintaining the chart only for Grafana Enterprise Loki users.
The project maintains connections with other major open-source ecosystems, with GitGenius identifying overlapping contributors with microsoft/vscode, microsoft/typescript, and rust-lang/rust. The codebase is distributed under AGPL-3.0-only with Apache-2.0 exceptions documented in LICENSING.md. The project provides multiple support channels including a community forum, Slack workspace, GitHub issues, and a Google Groups mailing list, reflecting its position as an established cloud-native observability tool within the Grafana ecosystem.
