By default only one instance of your application is ran, therefore there is nothing to load balance on, but if you choose to run more than 1 replica we will indeed load balance between the replicas via a randomized algorithm.

We do not support k8s, docker-compose, etc. Nor do we run users workloads with any of those or similar technologies.

I understand that you don't support containers but Do you support Nginx? And Can I increase the replicas on-air if I see that my application is phasing more load than usual if yes, then How do I increase the replicas? Will it restrart my application or in short term, will my users be effected in increasing replicas?

We do support containers, just not container orchestrators, as we have designed our own internally that is not exposed to the end user.

Yes you can run NGINX, but for the large majority of use cases that's not necessary nor recommend, and we handle SSL and HSTS for you.

Yes you can increase your replicas on the fly from within the service settings, this will redeploy but if you have your service properly setup with a readiness health check there is a zero downtime switch over for the new replicas when they come online, so as long as your have your service setup property your user's won't know anything.

Okay sounds good and best for next deployment, but this was a bad decision to not letting users to create templates for Free accounts, 5$ was already enough and now this limitation is bad as is and now also this github nag about verifying account to do anything which doesn't work always like if new account blah...blah...blah...

I'd have to disagree there unfortunately, it is done as an abuse prevention measure.

Unfortunately your GitHub did not provide enough information to verify yourself so you will have to upgrade to the Hobby plan.

Once on the Hobby plan there are no GitHub requirements since verification is instead done with the card that would be attached to the account.