Should we make a Certbot group?

A couple times now I’ve been asked if there is a Certbot group people can tag to get the attention of the Certbot developers rather than just pinging certain individuals every time something comes up.

So far my response has been that it would actually work better for us if people created a GitHub issue at https://github.com/certbot/certbot/issues/new. We regularly check that and share that work across the team so that would work quite well for us.

To be clear, I’m not advocating that most Certbot users should instead for help on our GitHub instead of this forum or that when a GitHub issue is created in response to a forum thread that the discussion should be moved there. It’s just the most convenient way for us for people experienced at answering questions on this forum to get our attention.

With that said, while I’d encourage people to try it and see how it goes, I’d like to know if this is too inconvenient for you all and in that case I can look into getting something like a Certbot developer group set up that people can tag.

5 Likes

There are the two extremes which are obvious.
The problem here lies in that great big grey area in the middle.
Like when a user doesn’t have a clue as to why things fail and the forum frontline support also has no concrete answer to their problem.
Should all those users take such problems directly to github?
[would the average user be able to maneuver github?]
Or should they simply be nudged up a tier from that frontline (tier 1) type support?
[maybe keeping the new group accessible to only a certain level users or higher would help]
The main question now (for me): Is there enough actual need for tier 2, 3, specialized typed groups?
[a benefit seems to be there]

1 Like

I, for one, definitely feel that the order of operations is clear from the perspective of those seeking advice, as evidenced by the very words of the OP of the most recent topic that triggered this discussion:

Granted, from my experience I’ve found that it’s usually edge-cases that are of the more complex sort that lead to bug investigations in the first place. This is entirely expected as the software equivalent of the combination of the adages “the road less traveled” and “the squeaky wheel gets the grease” naturally tends to be “the least used code paths yield the most issues”. Major issues with common paths tend to be rare due to excessive field testing and when they do arise they are quickly met with overwhelming force. Regardless though, I do not wish for patients with rare complaints to check themselves in through the service entrance of the emergency room and head straight to the surgical wing without going through the triage process just because it might be more expedient. At scale such a practice tends to lead to chaos, overburdens the crisis teams, and cripples efficiency. Some people will still do it anyhow, given the availability, but those people are usually more proficient than your average bear and thus less likely to be burdensome and can potentially add resources to those behind the curtain.

So, my recommendations are multifaceted:

  • Create the “certbotdev” group in the community for use by those providing assistance who lack the experience, comfort, or time to carry the issues directly over to GitHub. There will be some “patients” who will try to use this directly, but they will (hopefully) be well-informed patients. This is the “I need a doctor” call that requests triage by a physician rather than a nurse. This is similar to how the staff group currently operates.
  • Those of us familiar and comfortable enough with our specific community resources can still, of course, call out in certain situations for clarity of a situation that may not warrant the creation of an issue on GitHub, but would benefit from the response of a “physician” with certain experience. It looks good to both the members and newcomers of the community to see your smiling faces. :slightly_smiling_face:
  • Those of us members/regulars/leaders/staff proficient and confident enough to triage for ourselves and carry an issue over to GitHub (being sure to cross-link between the two communities) can aid the certbot development team by bringing direct information and awareness with efficiency. This is the experienced nurse who isn’t going to call a doctor and wait because it’s bleedingly obvious to them that the patient needs to go into the exam room stat. Let the doctor apply immediate efforts within the operations and system of the back of the house rather than having to wheel the patient around and transfer all the paperwork.

Anyhow, that’s my metaphor. I’ve created an account on GitHub aptly-named LetsEncryptCommunity. Ready for service.

2 Likes

Tangent: I think a dedicated “Certbot Help” topic might be helpful too.

1 Like

Isn’t that #help, generally?

You meant topic. Sorry, I read category.

I’m in the process of drafting a certbot/certification handbook:

https://community.letsencrypt.org/t/preliminary-discussions-for-a-handbook-of-certbot-best-practices/133296/39

You read my intent correctly.

Certbot has always been the most popular client in general, and on this forum. When LetsEncrypt first started, there were not many clients – but that has drastically changed.

The amount of people who are capable of helping with Certbot oriented questions is far greater than those who can help with the other clients.

I’m happy to spend 10 minutes to answer a question or two when I occasionally pop around here to find information on the (increasingly esoteric) features of LetsEncrypt/ACME I need to maintain our infrastructure and client… but the majority of posts involving Certbot don’t require much familiarity - they’re often from people who did not read the documentation or who read severely outdated documentation.

Segmenting the Certbot questions out, would essentially be a first-level triage, and the generic “Help” forum would mostly require some sort of specialized knowledge - either of the less popular clients or the ACME/LetsEncrypt API.

1 Like

An interesting proposal with some definite ramifications. With #help being the general category with massive internal and external linking, I’m not sure if moving what I would guess is 95%+ of the traffic to a new category is feasible or advisable. Most people will almost undoubtedly post in #help regardless of what forks we try, simply because it’s obvious. The nature of the people is hard to shift. It would probably be easier and more effective to create topic(s) for the other clients (especially the specialized ones) and either hope the public catches-on or spend the minutes a day needed to moderate the topics into the other categories to sort them out. Make me a mod and I’m happy to do it. :slightly_smiling_face: