[GitHub] [dolphinscheduler] rickchengx commented on issue #13017: [Feature][Remote Logging] Add support for writing task logs to remote storage

[GitBox <gi...@apache.org>]

rickchengx commented on issue #13017:
URL: https://github.com/apache/dolphinscheduler/issues/13017#issuecomment-1328731440

   > Good idea! One humble suggestion that **different log level** contributes to failure diagnosis differently, level of `error`, `warn`(high level) may contribute more than `info`, `debug`, `trace`(low level). Intuitively, it should better **keep more higher level log** and clean lower level log first when reach the maximum capacity of remote storage. Maybe we can separate different level of logs when flushing and support [log aggregation](https://bambrow.com/20221021-yarn-log-aggregation/#more) when reading logs like Yarn. WDYT?
   
   Hi, @Radeity , thanks a lot for your comment and suggestion.
   
   * This PR focus on writing task logs to remote storage after the task completes (E.g., it put the task log file to OSS). Also it reads the task log from the remote storage if the task log file does not exist on the local file system in worker.
   * So I think we should **keep the content consistency** of local logs and remote logs (E.g., the `log-2` and `log-3` in the image below) instead of aggregating logs on remote storage.
   * If the logs are aggregated and cleared according to the log level, users may be confused that he can only see part of the task logs.
   * So I think it is reasonable to clean up the oldest logs when the logs stored remotely reach the upper limit, just like `RollingFileAppender` in `logback`.
   
   <img width="630" alt="截屏2022-11-28 16 33 04" src="https://user-images.githubusercontent.com/38122586/204231822-b411664f-73c8-45a8-999a-397f85de8a71.png">
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@dolphinscheduler.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org