[dkersten]

I've interacted with teams that would commonly get messages in their slack channel asking to review PR's or whatever and the team would simply point at the channel status which was a link to their ticket submission page and the expected turnaround time, with a message like "Hi, please open a ticket and we'll get to it within the SLO", and then they strictly stuck to that. They also always linked the documentation if it might be in there, but with the note that if its not covered in the docs, to let them know. Usually it was in the docs.

Sure, people don't like it, but after a few attempts to bypass the system, they learn and stop doing it. (This is in a software company)

[alexchamberlain]

I really appreciate when someone takes a few minutes to answer my question on a public (rather than 1-1 chat), so I endeavour to do the same for others. Treat others as you'd like to be treated.

[dkersten]

Well, in this particular company, everything was public. Even the 1-1 discussions tended to be threads in the public slack channel. Or at least comments on the ticket.

The main point was if you want the other team to do something, you create a ticket so that they can schedule it in, rather than interrupting them demanding they drop whatever they might be doing to do your thing. And if your question is already addressed in the documentation, then its a waste of the teams time and energy to answer the same thing over and over.

If you have a legitimate question that isn't answered in the documentation, then you can absolutely ask and get an answer. The main goal is to reduce interruption with needlessly (everyone is busy!) and to reduce low quality low effort questions. Answers then often also made it back into the documentation.

So its not about treating the requester badly, but rather that its easy and low effort to ask someone to do or answer something, while it can be incredibly disruptive for the team being asked. This is a means of managing that, and in my opinion it worked great. It also pushed me to put enough effort into questions when I did have them (what am I trying to achieve, what have I tried, where have I looked for answers, what problems am I having).

[alexchamberlain]

This is why you should ask on a public chat; if I'm really too busy to read the Slack channel, I won't read the Slack channel. If, on the other hand, I need a breather or "I'm waiting for the compiler", I might be able to give you some pointers in the right direction or encouragement with what you've already discovered.

I'm not encouraging lazy questions per se; I want people to demonstrate effort (what have you tried?) and ask the right question (X-Y questions), but equally, I appreciate that documentation takes a long time to digest and maybe my opinion is what you want rather than discrete facts.

[AtlasBarfed]

Ticket walls are an unfortunate outgrowth of continued management incompetence in software development.

Unfortunately, once it starts it spreads like the plague, and combined with defensive internal service monopoly empire building, results in anything being done taking 10x more as a task passes through a half dozen first layer and who-knows-how-many second layer ticket walls/SLOs.

[malwrar]

What competence would make ticket walls go away? I’ve only ever had dayjobs with large orgs and ticket walls have been a useful defense against folks who insist on sending low-quality help emails or DMs the moment they encounter a problem. Sometimes I consider taking a massive paycut to work in a smaller company specifically to lessen the probability of needing to work with those sorts of people. If there’s some alternative model for dealing with these people I’m excited to hear about it.

[AtlasBarfed]

Aligning resources predictively to the needs of applications/projects. So if a project is about to go into intensive database development, you align a group of your DBAs to them for "preferential access".

Not just viewing the workers as completely interchangeable cogs, but getting them insight into specific applications, so requests can be processed more quickly.

Transparency into who is handling a request.

Managers scheduling and accepting enough "float" or unplanned work time to accommodate requests.

Educating developers to plan out their resources and requests.

Aligning support personnel to release schedules.

[phkahler]

>> What competence would make ticket walls go away?

I dont mean this as snark but how about making quality software that is easy to use and has good documentation.

More jokingly: Take down the walls and let some users actually bother developers with their problems until the developers fix those recurring annoyances just for their own sanity. This is probably going too far.

[dkersten]

The "ticket wall" I was talking about in my original comment had nothing to do with quality software, but rather getting a team handling a different competency to do or answer something. The example I gave was actually for the "cloud automation" team, they handled such things as AIM roles/policies, cross-account AWS permissions and various other AWS/cloud services that other teams used. Other teams had similar "go through tickets" ways to interface with them, for example the team that handled the servers that CI got provisioned on. Developer-services teams that provide services to development teams internally, basically.

So making quality software would not eliminate the need for them or the need for them to manage incoming requests.

[xapata]

Don't change jobs for that. Same problem everywhere.

[0xbadcafebee]

'Ticket walls' can be used as efficient pipelines for work. The problem is the teams who are asking for tickets aren't managing them correctly or don't have a good working agreement with the users. This is unfortunately very common with internal service providers.

But it can be changed. The users have to have a conversation with the service provider about what's wrong with the process, how it's affecting the users (& the business), and how they'd like it to change. If the service provider stonewalls them, they can go up the management chain. Often upper management has no idea what's going on and will get something done if they hear enough people complain.

[dkersten]

I liked it though because you could track the progress on the ticket instead of pinging people periodically to see what the status is. Once I got used to it, I found it was a much simpler way to get things done across teams.

[jasonpeacock]

What's the alternative to ticket walls? What would competent management look like in this case?