Logstash pipeline with S3 output starts tens of minutes
Nov 27, 2019
What to do when your logstash pipelines starts tens of minutes
When your logstash pipelines starts tens of minutes, you can have these problems:
- Not enought entropy - logstash doesn’t have enought random numbers
- You have a lot S3 outputs (doesn’t matter if one bucket or many of them Github issue)
S3 output plugin
Try to limit number of outputs. If you want to save logs to different paths in you S3 bucket. You can use mutate filter and add path to event. And then use only one S3 output.
Second option is to disable validation and metrics, that helps to speed start time.
validate_credentials_on_root_bucket=false
enable_metrics=false
Links
- Github issue
- Elasic docs for S3 output plugin
- How to debug Logstash
- Not enought entropy
- How to setup filebeat/logstash incredible tutorial