#169
A
Rank #23
Loki
Horizontally scalable log aggregation system designed like Prometheus but for logs — it indexes only metadata, keeping storage costs low by writing compressed log chunks to object storage.
80.0
/ 100
⚙ Full Stack
Go microservices (distributor, ingester, querier, compactor) that can run as a single monolithic binary for small deployments or split into independently scalable components for large ones, backed by object storage (S3-compatible) for chunks and an index.
📈 Scaling Analysis
Loki's index-light, object-storage-backed design is explicitly built for internet scale — the distributor/ingester/querier split allows each stage to scale independently, and the S3 backend means storage growth is effectively unbounded. Single-binary mode trades this scalability for simplicity and is the right choice until log volume genuinely demands the split.
🚀 Running on Nexlayer
For small-to-medium deployments, run Loki in single-binary mode as one pod plus an S3-compatible storage bucket; scale up to microservices mode (separate distributor/ingester/querier pods) only when volume demands it, wiring components via loki-distributor.pod:3100 style DNS. Mirror the official image via mirror.gcr.io and pair with Grafana or Promtail/Alloy for querying and shipping logs respectively.