Peerboard, Peercoin node statistics

[unvoid]

I'm developing an app on top of peercoin daemon that will show node operators statistics about their node in more user friendly manner rather than searching for info in CLI. Screenshot shows how it looks now. I'm kind of stuck now, don't have idea what can I put there to be more useful. Would be awesome to read your feedback/ideas.

[img width=950]http://i.imgur.com/BGfa416.png[/img]

EDIT: update screenshot

[peerchemist]

Excellent.

How is it implemented? What tools did you use and can you explain the architecture in few short lines?
I was looking for something like it for Peerbox, we can cooperate if you are interested.

[unvoid]

It's written in PHP (Symfony2). Using LEMP (could be LAMP of course) as host env and AdminLTE2 as UI. I'll write a good HOWTO (describing server configuration) once I release it to public.

It is now compatible only with Peercoin-QT and Peerunity. No support for peerbox currently.

Sure I would be glad if someone will give his time to help with this soft. Soon I'll make source public on GitHub.

[peerchemist]

I was hoping it is something more minimal and self hosted. Like a simple python-flask app or something implemented in Go.
I am kinda perfectionist when it comes to Peerbox, minimalist too. This application is just too heavy for me to include it in Peerbox.
Good luck though, be sure to open the code - I am sure people will find it useful.

[unvoid]

I'm trying as good as I can to make this app low resource requirements. Actually right now its consuming less than 10MB as a peak per request in development environment (in production it'll use less than that). The biggest threat to memory is DB engine (MariaDB/MySQL) but it could be tweaked by some tutorial to low levels as well.

Unfortunately I don't have any rpi to test it out in this env so I'll need to take feedback from any rpi users that will decide to run it.

[peerchemist]

The biggest threat to memory is DB engine (MariaDB/MySQL) but it could be tweaked by some tutorial to low levels as well.

Exactly, there is just too many dependencies for this application. DB should not be required.

[unvoid]

Initially I wanted to implement sqlite. This approach would be good if app would store only few hundreds of rows (i.e. data only from last week). But in the end I want it to be able to show statistics also in 3, 6 and 12 months periods so sqlite would die in those conditions.

But it's fine, after some tests maybe we could switch to sqlite. I want this app to be open for ideas and other people work.

[hrobeers]

Why would sqlite die?
I don't believe it is that much data.
I have the feeling that everything could even be held in memory.

[unvoid]

As I said in previous post at some point we can test out sqlite (if there will be enough interest from you guys). Right now I don't have a lot of faith in sqlite so I'll stay with MariaDB for now.

I deployed this soft to production server so you can look at it closer.

Full stack (Peercoin-QT, Nginx, PHP-FPM, MariaDB) using 440MB (with other basic soft like TOR, fail2ban, rkhunter etc). So in current stage it fit in rpi 1 if we speak about RAM (don't know how with CPU performance).

http://31.178.239.144/

[hrobeers]

Hi unvoid,

This is definately great work! We're not criticizing. Please continue the great work.

We are just pointing out a pain point that will impede wide adoption.

So please continue your great work and we'll follow with great interest.

Is it open source?

[CoinGame]

This is really freaking cool.

[Cybnate]

Nice work, I like this a lot.

Here is another idea; instead of a link to wikipedia when clicking on Connected peers show a list with IP addresses of connected peers and the protocol and client information they are on. Either as text or in a nice diagram.

Hope you can trim and perfect the solution as suggested by a few posters. Keep up the good work.

Maybe you can list it as a project on: http://peer4commit.com/projects. This is probably worth a few donations to encourage you continuing to work on this

[unvoid]

Thanks for few nice words and suggestions.

I've already thought about listing all connected peers but this feature is little more engaging because I would like to implement it with a information from which country peer is and display it on the map.

I have database of IPs with 118,000 records and still thinking if this will not be too much for most of hardware that people have here (I don't want to use some kind of external geoloc APIs because I think it's a security risk).

So I'll probably make a switch for those who don't want geoloc.

But yes, I'll try to find time to implement just list of peers first.

And thanks for the peer4commit link. I've created project there.