There have been some changes made to our code base that may cause you to encounter a 404 depending on your infrastructure/environment configuration for your SMP(self-managed-platform), previously known as on-premise installation.
The issue is straightforward, documentation has you navigate to:
This will allow you to create your initial harness admin account, however some users depending on security related measures taken in their environment, or if they have a type of ‘split-horizon’ DNS configuration where the name resolves to an internal address on the local network and external otherwise this can create the following issue seen in the developer console when signing up(F12, select network tab and look for the following):
The problem originates typically due to harness being accessed via an IP address that the load balancer used for the hostname did not initially resolve to(eg; the LB cannot be accessed from two IP’s currently due to design limitations). Be aware that the root cause can be different on a case by case basis, for me, the resolution was a simple DNS host record change.
I was accessing harness from my internal network via host A record in a local dns server as:
I updated this to my known external IP assigned to the load balancer(both would in fact resolve to the proper load balancer) but updating to the routable IP (non-rfc1918) addressed the issue and I was able to sign up. We have had identical problems occur due to egress traffic not being permitted to the load balancer IP assigned(external) as well in another example, so security policy can also play a big part here.
We are actively correcting this to rely on Kubernetes internal svc names between pods/services, which will speed things up and make this issue disappear hopefully very soon, thanks!