After we used graylog to collect logs, after the arrival of a large number of logs, the logs within 1 millisecond were printed too fast, and the timestamp time was the same after arriving on the platform. Therefore, we could not determine the order of logs for a period of time, we used the following method:
Secondary development of graylog with nanoseconds in the log, but the results are not satisfactory.
There are no good methods to use many acquisition agents, such as nxlog, filebeat, td-agent and logstash
Please help me，thanks