promtail basic config causes really high CPU load

[Danny de Waard]

I'm having trouble with using promtail in the most basic way.
When starting with only a few logfiles to monitor it causes kswapd0 to explode and my CPU-load goes to 50-60 (1m)

config:

server:

  http_listen_port: 9080

  grpc_listen_port: 0

positions:

  filename: /tmp/positions.yaml

clients:

  - url: http://myserver.com:3100/loki/api/v1/push

scrape_configs:

- job_name: system

  static_configs:

  - targets:

      - localhost

    labels:

      job: varlogs

      __path__: /var/log/*log

- job_name: x_server

  static_configs:

  - targets:

      - localhost

    labels:

      job: x_server

      __path__: /dir/server.log

- job_name: y_server

  static_configs:

  - targets:

      - localhost

    labels:

      job: y_server

      __path__: /dir/server.log

promtail version:

promtail, version 2.2.1 (branch: HEAD, revision: babea82e)

  build user:       root@e2d295b84e26

  build date:       2021-04-06T00:52:41Z

  go version:       go1.15.3

  platform:         linux/amd64

[Brian Candler]

I'm afraid you're asking on the wrong mailing list.  promtail is nothing to do with prometheus - it's part of the Grafana Loki project, which has its own discussion forum at https://community.grafana.com/