How did twitter generate the 1% of the Firehose tweets to status/filter endpoint and status/sample endpoint?



Could someone briefly tell me the algorithm? I have found this Topic is helpful:

However, it just answered part of my question. I need more details, especially deal with the time, thanks in advance,


As far as I know, 1% Sample Stream is based on Twitter IDs:

Snowflake IDs are based on a few things other than just a timestamp - and maybe changed over time, but there’s an old post I found useful:

But remember that this post was from 2013, and things may have changed considerably since then:


Thanks for your reply, that somehow help me:grinning: