Introducing Loki

Grafana Loki was designed from the ground up to be a highly scalable multi-tenant logging solution. Its design was heavily influenced by Prometheus with a few main objectives:

• It was built with developers and operators in mind (such as Diego and Ophelia, who were introduced in Chapter 1)
• It has simple ingestion; no pre-parsing is required
• It only indexes metadata about logs
• It stores everything in an object store

Let’s look at how Loki ingests data and uses labels as this will provide valuable insight into the way your queries source and then process the data for presentation:

• Log ingest: Loki accepts logs from all sources with a wide choice of agents available to make that easy. You can even send log data directly to the Loki API. This makes it the perfect choice for complex environments featuring a multitude of systems and hardware components.

  Loki stores its logs as log streams, where each entry has the following:

  • Timestamp...