I have a few crates published on crates.io which get mostly used by myself and thus have only a few daily downloads (0-200 depending on crate and day). These are mostly due to CI/CD workflows that test the packages.
I have noticed a change in the counter after the 30th of October along a few things:
Lower overall downloads despite no change in usage
No downloads of versions other than the most recent one.
You can verify these results for yourself by visiting:
I mean ... I get that spikes happen due to automation tools, possibly some ML companies who update their codebases to train their models on. But how does the inverse happen if usage does not change significantly. Was there a change in reporting of downloads on the side of crates.io?