It's not a functional back-up. Not for any commercial use case relevant outside the military.
If all you want is a back-up for military launches, the Ariane 5ME was a better, cheaper option that could have bridged the gap to a competitive reusable [1]. The billions of dollars wasted on Ariane 6 would have put Europe into the running for a competitive launch vehicle in the 2030s. Instead, we have Arianespace's CEO pitching another boondoggle to ensure Europe has a Falcon Heavy by the 2030s.
So yes, having 10% launch capability is better than zero. But that 10% could have been bought for much cheaper. And saying 90% of your space industry is subject to foreign control versus 100% is a bit milquetoast, particularly when the alternative would have been R&D to bring that down to e.g. 50%.
If ESA has 100 payloads, it can just book 90 of them on SpaceX and the mandated 10 of ArianeSpace (to keep the political pork happy), and lose only 10% efficiency (considering Ariane flight is twice as expensive as SpaceX).
The problem is that ESA has no scalable payload economy, nobody has a good reason yet to launch that much mass. SpaceX is its own customer with Starlink for scalable launches, but ESA or NASA will not deliver 100 sattelites per year.