gchristensen changed the topic of #nixos-borg to: https://www.patreon.com/ofborg https://monitoring.nix.ci/dashboard/db/ofborg?refresh=10s&orgId=1&from=now-1h&to=now "I get to skip reviewing the PHP code and just wait until it is rewritten in something sane, like POSIX shell. || https://logs.nix.samueldr.com/nixos-borg
orivej has quit [Ping timeout: 276 seconds]
orivej has joined #nixos-borg
orivej has quit [Read error: Connection reset by peer]
orivej has joined #nixos-borg
<LnL> grr
<gchristensen> so, this one didn't predict failure: https://nix.ci/prometheus/graph?g0.range_input=2w&g0.expr=predict_linear(node_filesystem_free%7Bmountpoint%3D%22%2F%22%7D%5B8h%5D%2C%2024%20*%203600)&g0.tab=0
<gchristensen> but also probably just time to reboot the machine :)
<mic921> I am also playing with netdata a bit: http://eve.r:19999/#menu_system_submenu_load;theme=slate;help=true
<mic921> Requires much less configuration to show useful stats. It only required a bit love to play well with nixos. Hopefully I can get this stuff upstream in the next time
<mic921> This link makes more sense: https://screenshots.firefox.com/KAUJGuYGN64YgtJS/eve.r
<gchristensen> not sure I can resolve eve.r :)
<gchristensen> ooh nice
<mic921> yeah I forgot this is my internal vpn domain.
<gchristensen> nice, how do you have that set up?
<mic921> It is a tinc mesh net I share with friends: https://github.com/Mic92/dotfiles/blob/master/nixos/vms/modules/retiolum.nix#L44 They have some proper way of generating the hosts list.
<gchristensen> hmm really cool
<mic921> But it is quite useful because I can get direct connections through firewalls.
<gchristensen> yeah
<gchristensen> I lose access to a lot of stuff when I leave the house, and I'm traveling more than ever so that is a bit annoying.
<mic921> I guess you can just join the network :) https://cgit.krebsco.de/stockholm/tree/krebs/3modules/lass/default.nix
* gchristensen reads up on tinc
<mic921> decentral, either layer2 or layer3, has internal routing when the other peer cannot be reached directly.
<mic921> Supports udp hole punching.
<gchristensen> this all sounds pretty good :)
<mic921> I thought I would switch to wireguard at some point, which is why I added it to systemd-networkd, I really need this udp hole punching so I stayed with tinc.
<mic921> *but I really
<ekleog> mic921: does tinc's hole punching actually work? my last experiments with ping didn't suggest it :/
<mic921> ekleog (IRC): To be honest the peers I tested it with had public ips but a stateful firewall. There it worked out for me. It might require at least one peer not to be behind a NAT or with port forwarding.
<mic921> We have like 3-4 public endpoints.
<ekleog> hmm yeah I've got a public endpoint, I guess the issue was both peers have private IPs
<ekleog> (I mean, the public endpoint is required so that I can get both ends to communicate anyway :°)
<mic921> In theory this should be possible. No idea.
<ekleog> hmm actually nvm it looks like now it works :D likely some testing error
<ekleog> (tested on a big download and looking at bmon on all 3 machines :°)
<LnL> mic921: netdata is nice, but it's pretty inflexible compared to a proper setup like prometheus
<mic921> LnL (IRC): I can imagine that, but it scales better down if one does not want to spend too much time on getting everything set up.
<mic921> For prometheus I need an exporter for everthing.
<LnL> for any system that's more than a couple of machines you'll pay the price at some point for not spending time on things like that
<mic921> Let's see. It can also stream data to a central instance.
<mic921> And in the end it can also just serve as data source for prometheus: https://github.com/netdata/netdata/wiki/Using-Netdata-with-Prometheus
<gchristensen> oh cool
<LnL> and how do you get your own metrics in there?
<LnL> oh the other way around that also works, but I'm not sure if it provides much more info than the node exporter
<mic921> You can write them in different languages, python seems to most popular one: https://github.com/coraxx/netdata_ts3_plugin
<mic921> The list of backends supported by default is already big: https://github.com/netdata/netdata/tree/master/collectors/python.d.plugin
<mic921> In most cases it tries to auto-detect what is there.
<mic921> In the end there are probably many ways to achive the same.
<LnL> hmm, guess you have more flexibility than I thought
<mic921> Our nixos-integration is a bit messed up.
<mic921> Basically all python plugins do not work. I have some fixes in the queue.
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos-borg
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos-borg
orivej has quit [Ping timeout: 252 seconds]
orivej has joined #nixos-borg
orivej has quit [Ping timeout: 252 seconds]
orivej has joined #nixos-borg
orivej has quit [Ping timeout: 245 seconds]
MichaelRaskin has joined #nixos-borg
orivej has joined #nixos-borg
<gchristensen> lol.
<gchristensen> LnL: let's try the 24hr / 24hr monitor
<LnL> yeah...
<{^_^}> [ofborg] @c0bw3b opened pull request #253 → Add c0bw3b to extra-known-users → https://git.io/fxg3n
<{^_^}> [ofborg] @c0bw3b closed pull request #253 → Add c0bw3b to extra-known-users → https://git.io/fxg3n