This is very unusual; if you can reproduce this problem, can you please follow the instructions here: https://dev.chromium.org/for-testers/providing-network-details to collect a network capture showing the failure?

Thanks!

Hello

Please find the attached log.

Deleted: net-internals-log (1).json 497 KB

net-internals-log (1).json 497 KB View Download

Hrm. This line number pointers here:
https://boringssl.googlesource.com/boringssl/+/54092ffeaa80ed032a50b914f14e3aa41de763df/ssl/tls_record.c#206

So somehow we're getting back something that's not even remotely a TLS record...

1. Would you be willing to provide a byte-level capture (see the bottom of https://dev.chromium.org/for-testers/providing-network-details)? That would let us see what's getting sent.

2. To clarify, when you say "Firefox:FAIL", do you mean the connection doesn't work in Firefox either?

3. Do you have some kind of antivirus or firewall or other networking product on your machine or network?

Thanks!

1. Would you be willing to provide a byte-level capture (see the bottom of https://dev.chromium.org/for-testers/providing-network-details)? That would let us see what's getting sent.

>> Unfortunately no. 

2. To clarify, when you say "Firefox:FAIL", do you mean the connection doesn't work in Firefox either?

>> Yes. "An error occurred during a connection to translate.googleapis.com. SSL received a record that exceeded the maximum permissible length. Error code: SSL_ERROR_RX_RECORD_TOO_LONG"

3. Do you have some kind of antivirus or firewall or other networking product on your machine or network?

>>> Yes. We have a firewall

What product is the firewall? That both Firefox and Chrome are unhappy and giving record-level errors strongly suggests to me that this is the firewall's fault and it is sending back some garbage rather than an actual TLS record.

I found the issue using wireshark. It's my proxy filter doing the magic

Alright. In that case, I guess we can close this issue.