Open ly123-liu opened 2 years ago
limits_config:
readline_rate: 4000
readline_burst: 8000
readline_rate_enabled: true
readline_rate_drop: false
->
limits_config:
readline_rate: 10000
readline_burst: 20000
readline_rate_enabled: true
readline_rate_drop: false
join slack channel https://slack.grafana.com/
sorry for so long , thanks, tests above those config helps, but still show some problems:
when promtail consumer one topic, promtail speed can to 19k/s
while promtail consumer many many topic , speed can only to 12k/s, when i scale promtail to above 12 or 24 pod , sped can only to 13w/s(while promtail consume specific kafka topic can be 30w/s), seems increase prometail can not up speed, why is that?
loki version:2.6.1 promtail version : main release
we test promtail used as kafka consumer to loki, shows one promtail only to 4k/s speed(while use logstash can be 1w/s), promtail just can't impove speed anymore ,is my configuration not correct?
monitoring image:
promtail configuration: server: http_listen_address: 0.0.0.0 http_listen_port: 9080 grpc_server_max_recv_msg_size: 104857600 grpc_server_max_send_msg_size: 104857600
positions: filename: /promtail/positions.yaml
clients:
scrape_configs:
- action: replace
source_labels:
- __meta_kafka_partition
target_label: partition
- action: replace
source_labels:
- __meta_kafka_group_id
target_label: group
limits_config: readline_rate: 4000 readline_burst: 8000 readline_rate_enabled: true readline_rate_drop: false max_streams: 0