Stop using pilot in foo.pilot.2i2c.cloud for new hubs

[choldgraf]

Description of problem and opportunity to address it

Context to understand the problem
Currently, the hubs on our shared cluster follow the pattern <hubname>.pilot.2i2c.cloud.

However, we are past our pilot phase, and calling it a pilot may cause organizations to make assumptions about the quality or reliability of our hubs. Moreover, pilot isn't particularly descriptive, it's more of an implementation details from 2i2c's perspective rather than something that a hub's user cares about.

Proposed solution
We should instead use the pattern <foo>.hub.2i2c.cloud. This will be:

• more memorable
• a clear indication that there's a hub at the URL
• less tied to the "phase of 2i2c's service"

Implementation guide and constraints

• We should probably set up a redirect so that *.2i2c.cloud redirects to *.hub.2i2c.cloud
• One reason we might not want to do this is if we explicitly want to have different hubs on different parts of our infrastructure (e.g. alpha.2i2c.cloud as well). But thus far I don't think we're doing this.
• As part of this, we should also rename github.com/2i2c-org/pilot to https://github.com/2i2c-org/user-guide or something like this.

Updates and ongoing work

No response

[choldgraf]

Could I get some feedback from the @2i2c-org/tech-team on this one? I don't think that making this change will be too difficult and the longer we keep it as pilot, the more annoyance there will be if/when we change it.

[yuvipanda]

Another option is to just move them to .2i2c.cloud.

[choldgraf]

@yuvipanda I'm +1 on that as well! I defer to the preference/judgment of others here, just think that we should move away from the word pilot

[sgibson91]

No pushback from me about removing pilot - I suspect no one has had the time to action this yet :D

[consideRatio]

Notes from sprint team meeting

@sgibson91 summarized that the remaining action points may be of a technical nature of updating domains pointing to hubs. @yuvipanda suggests that we could choose to not update these and just switch to ensure new hubs get modern domains.

The worry is that redirects could be troublesome to get right.

---

Conclusion:

• We opt to not setup redirects etc for old hubs at this point in time. We could in the future if we have a concrete technical plan on how to get it done.

---

@sgibson91 mentions that we would need to update callback urls as registered in the GitHub OAuth clients we have reigstered manually. This is relevant IF we change a domain name. Sarah mentions that this is not relevant for Auth0 OAuth clients, as they are automatically updated by the deployer script.

[choldgraf]

+1 from me on not renaming existing hubs unless they ask for it, and just start new hubs without pilot

[consideRatio]

@choldgraf is there any remaining action points besides this for this issue then?

[choldgraf]

@consideRatio I think it would be "documenting the new practice of not using pilot in our new hub setup docs, to formalize this plan from our team". Does that make sense?

[consideRatio]

@choldgraf can I assign you to fix that? I note that you have done a lot of work recently to migrate docs across the 2i2c org

[choldgraf]

@consideRatio sure thing 👍

Ah one question I thought up for redirects. What about:

• grafana.pilot.2i2c.cloud
• demo.pilot.2i2c.cloud
• staging.pilot.2i2c.cloud

or other hubs that are 2i2c-specific? Is that worth re-naming? I don't feel strongly but wanted to note it

[consideRatio]

@yuvipanda suggests that we can move grafana/staging and it will be fine. For demo we would need to come up with a practical solution on how to redirect. Also, what is a priority? This isn't seen as a priority among us currently, but it can change if we get input from users about it - for example about users concerned about this is a pilot project etc.

[choldgraf]

In my opinion demo.pilot.2i2c.org could move without worry about old links breaking, we have shared this in a very limited fashion.

I don't think this is a huge priority, but the main thing I am worrying about is the feedback I got from some communities that the word "pilot" conveys "unreliable and unstable"

[yuvipanda]

#1129 moves a few of these

[yuvipanda]

@choldgraf demo.2i2c.cloud works now! Can you tell the rocky mountain lab folks to use the new URL?

[choldgraf]

done! think we can close one?

[choldgraf]

I'm going to assume the answer is yes!