aderm commented on issue #4190: Recommendations for reducing the number of indexes URL: https://github.com/apache/skywalking/issues/4190#issuecomment-572362815 > > Deleting data is the same as before, but when deleting, add use type = hour or day or month to filter.This is the same as the previous logic. The only difference is the index name parameter, which is filtered by the type field. > > No, it doesn't. Currently, from 6.2.0, deleting mean remove the whole index, rather than data, which will cause elasticsearch GC or even FGC. So if you put different precision data in the same index, this is a problem. <img width="368" alt="WX20200109-110838@2x" src="https://user-images.githubusercontent.com/2892433/72034538-4591f080-32d0-11ea-8aa5-3b1a32a260bc.png";> I'm also talking about deleting the index, not deleting the data inside. Deleting data does have performance problems, which is related to the principle of inverted index. For example, all_p75_day-20191225 and all_p75_hour-20191225 are deleted on a daily basis, But they can be merged into one, and the entire index is deleted by day after the merge..
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
![https://user-images.githubusercontent.com/2892433/72034538-4591f080-32d0-11ea-8aa5-3b1a32a260bc.png"](https://user-images.githubusercontent.com/2892433/72034538-4591f080-32d0-11ea-8aa5-3b1a32a260bc.png"){kind=link}