The "no more god mode" (#2135) should be counted as a smaller must-have item. It should take a few hours at most unless I'm missing something about our implementation.

#2995 is worth consideration. It's probably amongst our top 3 feature requests in terms of feedback.

Note to carry over Resource Metadata and template.yaml from the July roadmap (#3042) that almost certainly will not be finished.

Is Frontend Edit Workspace (#802) really a big feature? The issue describes a page with simple forms and we already have the backend endpoints.

Is Frontend Edit Workspace really a big feature? The issue describes a page with simple forms and we already have the backend endpoints.

@ammario i think the misalignment is that we don't have a rubric for when something is big enough to be considered a "major" item on the roadmap. I think you're right, though. This one should not be on the "major" list.

@tjcran a heuristic I use is whether we would include the feature in a release blog post. There's still a lot of room for interpretation, but perhaps slightly less.

@tjcran @kylecarbs I feel like I'm seeing a pattern of bugs related to orphaned resources in the cloud provider. It's pretty concerning because it can cause wasted cloud spend. Our architecture mostly trusts the provisioner (terraform) to clean up after itself, but there are clearly edge cases like cancellation of jobs, bugs in terraform providers, etc. I wonder whether we need to think about some sort of reconciliation process.

@tjcran @kylecarbs I feel like I'm seeing a pattern of bugs related to orphaned resources in the cloud provider. It's pretty concerning because it can cause wasted cloud spend. Our architecture mostly trusts the provisioner (terraform) to clean up after itself, but there are clearly edge cases like cancellation of jobs, bugs in terraform providers, etc. I wonder whether we need to think about some sort of reconciliation process.

This sounds like it needs an epic to wrap all these together? Is there a need for an RFC on this to consider a more holistic strategy for dealing with it?

Also noting #3048 as related

I'd want to start with some strategic discussion & brainstorming, maybe @ketang @kylecarbs @spikecurtis @tjcran (optionally @ammario, optionally other engineers with deep distributed systems experience). One outcome of that discussion could be to move to RFC, but it could also be that we're going to punt on the larger problem and just fix bugs.

I'd want to start with some strategic discussion & brainstorming, maybe @ketang @kylecarbs @spikecurtis @tjcran (optionally @ammario, optionally other engineers with deep distributed systems experience). One outcome of that discussion could be to move to RFC, but it could also be that we're going to punt on the larger problem and just fix bugs.

@spikecurtis is this RFC something you are proposing for the August roadmap?

My personal feeling is that this trumps several things on the August list, like activity based auto-off.

My personal feeling is that this trumps several things on the August list, like activity based auto-off.

I have encountered the problem personally and I agree it's a fundamental problem with our model that we need to resolve. We have a few issues related to this problem, and no silver bullet has been identified. There are a bunch of small things that make progress, e.g #3035 and #2256 (comment).

But it's not clear what exactly we would pursue on if the "orphanage problem" was added to roadmap.

You can probably think of more/better solutions than the aforementioned issues too.

Another thing to note is activity based auto stop has been requested many times by users where orphanage resolution hasn't.

I suggest we add the orphaning problem to the roadmap as an RFC item for August so we can work out how to solve it. But I don't think we can properly assess the cost and prioritize it to this month's roadmap without an RFC.

Just something to be aware of: Licensing and Audit Logging aren't totally separable. They each have independent tasks to start with, but when it's time to add audit logging to the app, it's also time to gate audit logging by license. Seen from the other side, we can't fully implement licensing without a conditional feature to implement it on.

#1763 should be replaced by the first of the two tickets replacing it (the second is less urgent so it's up to you whether it makes it on the August roadmap)

I'm not clear on the requirements for #802 / #2660. #802 was going to be a page you can navigate to whenever you want to change params. #2660 sounds like it's asking for the app to check for template changes when the user clicks Update, and if they exist, display a form to fill in those inputs.

Just something to be aware of: Licensing and Audit Logging aren't totally separable. They each have independent tasks to start with, but when it's time to add audit logging to the app, it's also time to gate audit logging by license. Seen from the other side, we can't fully implement licensing without a conditional feature to implement it on.

#1763 should be replaced by the first of the two tickets replacing it (the second is less urgent so it's up to you whether it makes it on the August roadmap)

I'm not clear on the requirements for #802 / #2660. #802 was going to be a page you can navigate to whenever you want to change params. #2660 sounds like it's asking for the app to check for template changes when the user clicks Update, and if they exist, display a form to fill in those inputs.

@presleyp good callout. I'll make the update for #1763 item.
The workspace edit issue needs some prep work and some product direction. @bpmct were you still planning to work on that one?

@presleyp #802 and #2660 is something that we need to define more in August! However, it is still something we want to do in August. It needs an owner and I'd estimate it's fairly complex and needs backend work as well.

For enterprise, #3304 is a good simple initial feature and we've already had a prospect request it.

Here is another simple yet urgent issue since it will help us better understand this early usage and make a better roadmap in September.

Suggesting we add #2662 as well as a nice-to-have.

Edit: or even a must-have. A lot of our docs will run off scripts to install code-server or projector, and health checks help avoid broken experiences for users before the IDE loads.