-
Notifications
You must be signed in to change notification settings - Fork 3.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
InfluxDB should have static admin assets compiled into the binary #1032
Conversation
i ❤️ https://github.com/jteeuwen/go-bindata FWIW |
This may be a wild idea but why not just entirely remove the admin interface from the main demon ? I clearly don't share your excitement about embedding a bunch of static files into an executable, I don't use the admin interface (I just don't see any reason to use it right now) and having a bigger executable for nothing is not that great. At least if you do that please make it optional so I can continue building a "naked" version of the influxdb server. |
really, does this matter in 2014? can't imagine the size increase is significant compared to ram sizes. |
@schmurfy I think your first proposal is a direction we may take down the road, but for now I think we're choosing to prioritize getting up and running quickly over file size. The usability gains are pretty huge and the only clear downside that I can see is adding an extra megabyte or so to the executable. It also gives contributors an easy way to test out functionality locally without installing an extra package. As it was, having the assets outside of the repo caused enough of a headache, so I'd prefer to step closer to integration than further away. With that said, I feel that adding an optional build flag to exclude those assets is a good solution for those wishing to slim down a bit. For now, you should be able to do the following:
Then build as normal. Let me know how that strikes you. |
e3b2029
to
fdc9235
Compare
@toddboom Thanks, as long as it is optional it is fine for me. @Dieterbe you are right the size is not really the matter that's more a philosophical concern, I don't like dead weight especially in this case it feels like a mad scientist idea, I understand that for testing and development environment it might make things easier though but I just don't see the point in production. Don't take me wrong influxdb is awesome, I waited for a long time for such project to be able to drop RRD (and that's why I am following its development) but I think that opposing opinions on matters like this gain to be shared, in this case having it as optional is a good enough solution for me :) |
I think pragmatism and simplicity are core values in the influxdb project and this looks like a simple and pragmatic solution, so maybe that helps in understanding the reasoning @schmurfy |
by the way, what is benefit to serve static out of backend rather than web-server? From ops point of view this brings a certain level of complexity. To be not able to change static quickly and serve admin/api endpoint with sub-URI makes such great product difficult to deploy and maintain over time. How can we improve that? |
They are not targeting sysadmin with this but end users wanting to try quickly influxdb who just have to run an executable and it's done, I can understand that. |
So far, I ❤️ statik (https://github.com/rakyll/statik).
We'll probably need to make a couple tweaks to the build process, but it should be pretty straightforward since we're already building the static assets.