>ALL your services accessible through the tunnel are "down" for your users

Not all.

I operate site with IPv6 only origins behind cloudflare.

During the outage I manged to login to the dashboard after some time and remove cloudflare for nearly 2 hours, and traffic level stayed close to 50% during the IPv6 only period.

Nobody complained: those who did not have working IPv6 probably blamed it on cloudflare.

▲

dpark 5 hours ago | parent [-]

> traffic level stayed close to 50% during the IPv6 only period.

> Nobody complained: those who did not have working IPv6 probably blamed it on cloudflare.

You described a situation where the outage resulted in 50% of your customers were unable to reach you and you were unable to do anything about it. I don’t think this story is a win for IPv6, regardless of whether your customers blame CloudFlare or not.

▲

patmorgan23 31 minutes ago | parent | next [-]

This has nothing to do with anything inherent to IPv6 and everything to do with the failure of organizations to timely implement it.

	▲	dpark 28 minutes ago \| parent [-]
		I didn’t say it was an issue inherent to IPv6. But it is a practical issue with IPv6.

▲

orangeboats an hour ago | parent | prev [-]

Compared to 0% like others?

50% is a very substantial retention rate.

▲

dpark 33 minutes ago | parent [-]

Would hand been 100% if his site supported ipv4 natively instead of relying on CloudFlare to do the translation.

The story here is not “ipv6 made my site resilient to CloudFlare outage”. It’s “50% of my customers can’t reach my site even when I turn off CloudFlare”.

	▲	orangeboats 22 minutes ago \| parent [-]
		>if his site supported ipv4 natively And it's becoming difficult for people to do so precisely because of IPv4 addresses running out...