quic-dev, is there a known issue that's responsible for a larger percent of QUIC_PROTOCOL_ERRORs on Linux Stable around April 4th?

Variations.FailedRequestErrorCode - QUIC_PROTOCOL_ERROR (http://shortn/_5uD1UwId28) - is back to normal. It seems to chirp once a week (goes as high as 3% on linux, lot of variation).

On Windows (http://shortn/_sI1OPMA0aa), there was an increase Oct 29th (Thu) 2015 and it has increased slightly (120k vs 140k) and it is almost proportional to traffic increase.

On Android, we are getting lot more errors (relative to #of ActiveSessions)

On Wed, Apr 6, 2016 at 11:32 AM, Alexei Svitkine <asvitkine@google.com> wrote:
Well, if failures get too high then people aren't successfully getting Finch updates. In this case, it looks like this error code went from being responsible for ~1% of Finch fetch failures to 3.5% which seems significant enough that it's worth looping +quic-dev to it, just in case.

quic-dev, is there a known issue that's responsible for a larger percent of QUIC_PROTOCOL_ERRORs on Linux Stable around April 4th?