That’s great @uberannie, the Failed to fetch error log gives us a way forward.
I would try disabling/enabling git-gateway from the Netlify site to make sure it is not a permissions issue.
The next step would be to try and find the request that is failing from the browser network tab.
You might find it useful to check the preserve log option when doing so (screen shot from Chrome):
cool, let me try the disable / enable option. In the interim, it looks to be the .netlify/identity/settings and .netlify/identity/user requests that are failing
Note there are a couple of team members who have no problems. They are also using chrome. So I’m wondering if there is anything local storage or cached to be considered?
Hi @uberannie, clearing the cache can help reduce the scope and would be helpful.
Does the browser reports any CORS errors or any other error on why the request is canceled?
Ok I did a couple of things which resolved it, could these inform what is going on?
cleared local storage
instead of using our prod URL/admin, I used the netlify URL/admin
CMS loaded as expected
(interestingly, after it loaded successfully using netlify URL/admin, using prod URL/admin also works But going prod URL/admin first resulted in endless loop…
Question - should both /admin URLs work or should we only use the Netlify generated URL ?
Both URLs should work.
Do you have any warnings in the custom domain management section in Netlify?
I’m assuming curl -i "https://prod domain/.netlify/identity/settings" fails and curl -i "https://netlify domain/.netlify/identity/settings" succeeds?
Hi Erez, there’s only a warning regarding an apex domain.
Not sure if related but should the CMS work on branch deploys? ie [branch url]/admin ? I’m getting the following error - but everything seems ok so far re our prod site…
The CMS should work on branch deploys, that might be another issue (again we would need to look at the traffic to figure this out).
Please note that unless you configure the CMS to connect to a specific branch it will always connect to the default branch.
Meaning branch deploys and prod will connect to the default branch.
You can dynamically update your config.yml during build time to get the CMS to connect to the specific deploy branch.
backend:
name: git-gateway
branch: master
Netlify supplies a HEAD env variable you can use for that purpose.
OK so I think our prod “endless loop” issue might be OK I’ll have to check in with the rest of the team - I can’t replicate now so I’ll get them to clear cookies/local storage etc to see if that resolves their issue as it seems to have fixed it on my machine, for our prod site
Still seeing endless loop issue on our branch deploy, clearing cookies etc doesn’t seem to resolve I wasn’t involved in the setup so might have to do a bit of digging on how it was setup.
The curl returns a json payload which I imagine is OK
hey guys, sorry been a bit quiet as we’re busy busy working through some deadlines to release the next version of our public website
so we’re still having it intermittently and randomly amongst team members - some no issues, some every now and then.
was chatting with one of our engineers, is there anything in the codebase that has some dependencies between calls, as a race condition? the theory that one call is waiting for another but timing is messing things up. One suggestion was to apply a “slow 3G” network mode in Chrome. This has actually “fixed” the issue so wondering if the forced slower connection helps manage the orchestration of calls?
But going 
