What settings can I adjust to reduce memory usage when querying?
I’m running loki in a single instance on a small 1CPU node and limiting memory to only a few hundred MB.
I was thinking reducing parallelism somehow
frontend_worker:
parallelism: 1
Would reducing limits_config.max_query_parallelism also help?
My grafana dashboard isn’t able to complete all the loki queries without loki running out of memory.
My queries also load json twice (embedded json text) like the following which probably doesn’t help…
{app="ingress-nginx"} !~ "(grafana|jaeger)" | json | line_format "{{.log}}" | json