Problem statement

Currently our smallest LokiStack deployment size is 1x.small and targets ingestion loads up to 500GB per day. This is yields a resource requirements profile that is not applicable to users with very small ingestion volume (e.g. 30GB, 60GB or 100-200GB per day).

Goals

• Propose a resource requirements profile for a 100GB / day ingestion volume.
• Propose a resource requirements profile for a 250GB /day ingestion volume.
• Collect and document any limitations and trade-offs running Loki at that small scale.

Developer Notes

• 1x.extra-small can be evaluated indenpendently from the fact that it used to be a the demo setup. We currently have 1x.demo for the latter.
• The replication factor should stay at 2 to support a proper HA setup.
• The replica count per component should stay to 2 to support a proper HA setup.
• Evaluate impact on current EPICs in progress: LOG-3266, LOG-3839 and LOG-3840