Hello,
I'm trying to see if Secor helps solve an issue we're facing - basically knowing when something is safe to read from S3.
I started here -
One of the stated objectives here is:
low log-to-ready_for_consumption delay: logged messages should be ready for consumption by analytical tools asap.
Reading the design doc, it's not clear to me how this is supported.
In the uploader.check_policy() method, there's a call to upload_files_to_s3(t, p). Now, if we have an external reader, an analytics tool looking at S3 as a data source, how do we get around the eventual consistency issues here?
Perhaps I'm missing something, appreciate any pointers!
Thanks!
Neil