Grafana Loki Logging

Loki is an open-source log aggregation system created by Tom Wilkie and the Grafana Labs team in 2018, designed as a horizontally scalable, highly available, and cost-effective alternative to Elasticsearch for log management, inspired by the Prometheus architecture and using the same labeling system, storing only metadata about log lines rather than indexing the full text content, dramatically reducing storage costs and operational complexity while maintaining fast log querying through label-based filtering, integrated natively with Grafana for visualization, adopted by over ten thousand organizations including Squarespace, DigitalOcean, and GitHub for their logging needs. The label-based indexing approach means Loki indexes only the metadata labels associated with log streams rather than the full text content of each log line, resulting in storage requirements typically ten to one hundred times smaller than full-text indexing systems, with the label system using key-value pairs identical to Prometheus, enabling seamless correlation between metrics and logs through shared label sets, while the LogQL query language provides filtering, extraction, and aggregation of log data through a syntax familiar to PromQL users. The distributed architecture consists of multiple horizontally scalable microservices including distributors for ingesting log entries, ingesters for buffering and flushing to storage, queriers for executing queries, query-frontend for splitting and caching queries, rulers for alerting, and compactors for managing retention, with state stored in a key-value store and chunks stored in object storage including Amazon S3, Google Cloud Storage, Azure Blob Storage, and local filesystem. The Promtail agent collects logs from files and systemd journal. The multitenancy support. The alerting through Alertmanager. The native integration with Grafana dashboards. Designed by Grafana Labs. Designed for DevOps and operations teams.