Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Grafana Loki: 1 year later

Grafana Loki: 1 year later

Presentation deck for follow up Loki talk given from DevOpsDaysIndia 2019

Avatar for Sandeep Sukhani

Sandeep Sukhani

September 13, 2019
Tweet

Other Decks in Technology

Transcript

  1. Layout • Quick Loki Intro • New Features • Optimizations

    • Demo • Running Loki in Production • Community Support & Future
  2. 03/18 Project started 12/18 Launched at KubeCon 12/18 #1 on

    HN for ~12hrs! 03/19 Labels from logs 04/19 Lazy Queriers; LogQL filtering 05/19 Live Tailing 09/19 7k GitHub stars https://github.com/grafana/loki Loki is a horizontally-scalable, highly-available, multitenant log aggregation system inspired by Prometheus.
  3. LogQL • Log Stream Selector ◦ {app=”foo”, namespace=”dev”} ◦ {foo=~”.+bar”}

    • Filter Expression ◦ {app=”foo”, namespace=”dev”} |= “error” ◦ {app=”foo”, namespace=”dev”} |= “error” |= “timeout” ◦ {app=”foo”, namespace=”dev”} |= “error” != “timeout” • Aggregation operators ◦ rate({app=”foo”, namespace=”dev”} |= “error” [5m]) ◦ count_over_time({app=”foo”, namespace=”dev”} |= "error" [5m])
  4. # Optimizations Stats from internal Loki deployment: • ~17 MB/s

    log ingestion • Upto 100K entries per chunk • Upto 600KB chunk size • Upto 80K chunks fetched per query Areas of improvements • Memory leaks • Batch Processing & Memory Pools • Benchmarking • Monitoring
  5. # Running Loki in Production • Deployment Modes ◦ Micro-service

    vs Monolith • Deployment Options ◦ Binary ◦ Docker ◦ Helm ◦ ksonnet … • Scaling
  6. # Community Support & Future • Governance • Community Calls

    • Releases • New upcoming features • GA