BLER

BLER tells you what fraction of transport blocks arrived corrupted — failed their CRC — out of everything sent. It is one of the most useful health metrics on a link because it closes the loop on link adaptation. The scheduler isn't trying to drive BLER to zero, which sounds counterintuitive; it deliberately targets around 10% initial BLER, leaning on HARQ retransmissions to clean up the rest.

The reason is efficiency. Aiming for zero errors would mean choosing an over-conservative modulation and wasting spectrum on a link that could carry more. Sitting at roughly 10% means you are pushing the modulation as high as the channel allows and letting fast HARQ recover the stragglers. So when you see BLER climbing well above that operating point, it signals the link adaptation can't keep up — the channel is degrading faster than feedback can track it, often a sign of mobility, interference, or a coverage hole.