The potential culprit doesn't feel right. Adding the appropriate component, so the owners can take a look at it.

Network bug triager here.

I agree the identified culprit seems unrelated.  

The timing also doesn't obviously match changes to the active portal code.   jam@ your changes are fairly recent, do you have any thoughts on this?

[+malaykeshav]

So FindIt points at one CL, which is a ChromeOS change related to zoom settings.  The crash trace has OnResizeOrRepaintACK() in it, which is presumably related to window sizing, which is...also potentially related to zoom settings?  The crash also looks to be on "linux-chromeos-dbg", so ChromeOS.  I think FindIt may be right about this one?

#0 0x7fa6f89e154d base::debug::StackTrace::StackTrace()
#1 0x7fa6f89dfb7c base::debug::StackTrace::StackTrace()
#2 0x7fa6f8a60f3d logging::LogMessage::~LogMessage()
#3 0x7fa6f0ae6cb0 content::RenderWidgetHostImpl::OnResizeOrRepaintACK()
#4 0x7fa6efa297bd _ZN4base20DispatchToMethodImplIPN5media29GpuVideoDecodeAcceleratorHostEMS2_FvRK50AcceleratedVideoDecoderHostMsg_PictureReady_ParamsENSt3__15tupleIJS4_EEEJLm0EEEEvRKT_T0_OT1_NS9_16integer_sequenceImJXspT2_EEEE
#5 0x7fa6efa29718 _ZN4base16DispatchToMethodIPN5media29GpuVideoDecodeAcceleratorHostEMS2_FvRK50AcceleratedVideoDecoderHostMsg_PictureReady_ParamsENSt3__15tupleIJS4_EEEEEvRKT_T0_OT1_
#6 0x7fa6f0b00777 _ZN3IPC16DispatchToMethodIN7content20RenderWidgetHostImplEMS2_FvRK38ViewHostMsg_ResizeOrRepaint_ACK_ParamsEvNSt3__15tupleIJS3_EEEEEvPT_T0_PT1_OT2_
#7 0x7fa6f0afb551 _ZN3IPC8MessageTI36ViewHostMsg_ResizeOrRepaint_ACK_MetaNSt3__15tupleIJ38ViewHostMsg_ResizeOrRepaint_ACK_ParamsEEEvE8DispatchIN7content20RenderWidgetHostImplES9_vMS9_FvRKS4_EEEbPKNS_7MessageEPT_PT0_PT1_T2_
#8 0x7fa6f0ae5d8b content::RenderWidgetHostImpl::OnMessageReceived()
#9 0x7fa6f0a8e4b3 content::RenderProcessHostImpl::OnMessageReceived()

Also note that none of the captive portal tests resize tabs, though they do open / navigation tabs.  So new a CHECK in sizing code seems rather weird, and unlike to be due to the captive portal code itself, or the test, unless there's some weirdness around navigation.

The failure definitely does not seem to be related to the CaptivePortal test code. However I do not see a direct relation between my change and the CHECK failure in the log. 
Are there other tests becoming flaky with the same crash log?

I unfortunately have no idea - none of the links in the bug report contain links to anything (Including whatever system detected the failure).