What I love about Grafana Loki:
- deploy it as a single binary for easy testing,
- go full distributed/microservice mode for production
Via Helm: https://github.com/grafana/helm-charts/tree/main/charts/loki-distributed (thx @unguiculus )
Our like our production setup via Tanka https://github.com/grafana/loki/tree/master/production/ksonnet
#grafanaloki #prometheus
- deploy it as a single binary for easy testing,
- go full distributed/microservice mode for production
Via Helm: https://github.com/grafana/helm-charts/tree/main/charts/loki-distributed (thx @unguiculus )
Our like our production setup via Tanka https://github.com/grafana/loki/tree/master/production/ksonnet
#grafanaloki #prometheus
As you can see in the diagram, both read and write paths can be scaled independently.
What's also great is that the write path is incredibly stable and efficient.
Need more oomph on the query side? Add more queriers!
What's also great is that the write path is incredibly stable and efficient.
Need more oomph on the query side? Add more queriers!
And Grafana Loki v2+ uses 100% object store, just like #cortex for Prometheus and #tempo for Tracing.
Object storage is very cost-effective, so you could say that with Loki you:
"Pay for the needle, not the entire haystack" - @sfingerhut
Object storage is very cost-effective, so you could say that with Loki you:
"Pay for the needle, not the entire haystack" - @sfingerhut
where is that Twitter edit button! forgot to also tag @torsten_walter for the Loki Helm chart contribution