Loki
Loki is Grafana Labs’ horizontally scalable log aggregation system inspired by Prometheus labels. It indexes metadata, not full log text, and pairs with Grafana for log exploration. We trial it for Kubernetes and service logs alongside Prometheus metrics.
Blurb
Loki is a horizontally-scalable, highly-available, multi-tenant log aggregation system inspired by Prometheus.
Summary
What it is: Log ingest (Promtail, Alloy, or OTel), label-based indexing, LogQL queries in Grafana.
When to use: Grafana-centric observability stack; cost-aware log retention with object storage backends.
When to skip: Full-text SIEM requirements (Splunk, Elasticsearch). Teams without label discipline on logs.
Key features: LogQL, multi-tenancy, boltdb/TSDB schema evolution, Grafana correlations with metrics and traces.
Details
| Topic | Notes |
|---|---|
| Deploy | Helm on Kubernetes; microservices or simple scalable modes |
| Pairing | Default companion to Prometheus and Grafana in LGTM stacks |
References