It was quite a lot of #MastoAdmin today with @KayOhtie to figure out that the Vultr node his don't-call-it-a-droplet was provisioned on seems to be failing (probably storage, maybe RAM?). But in the process, it got updated nginx configs, a clean slate for the database (dump and restore), cleaned-out redis, and more. Plus, a fancy #Prometheus + #Grafana stats dashboard (thanks @IPngNetworks!) that I'm currently #selfhosting to keep an eye on perf. The new "droplet" is literally 10x faster in Postgres and Redis ops.
It also got a simple "failover" media proxy:
#Cloudflare to mediacdn.blimps.xyz (Nginx on the droplet) => Cloudflare-fronted B2; if B2 is 404, serve Cloudflare-fronted DO Spaces instead.
Should keep egress costs down from DO and zero from B2 via caching and B2+CF Bandwidth Alliance.
Finally got #pgwatch2 to monitor my lab #PostgreSQL#Databases up an running on #openSUSE Leap 15.4. It was quite a bit of work to figure out the setup and it was anything but easy.
What's the best web based monitoring system (CPU, RAM, disk, network, etc.)? I don't like Grafana (don't ask me what's wrong with it, I'm a guy with a precise taste for aesthetics). #MastoAdmin#Server#Linux
@rolle I use both #LibreNMS (linux-centric - SNMP or librenms agent only), and annoying to contribute to, and #NetXMS which is awesome, java based, easy to contribute to, but with a learning curve like this: / #Grafana is nice for graphics, but is no monitoring system...
I could demo either of them via #jitsi or similar sometime if you're interested...
Honestly #GrafanaCloud is a piece of crap, at least if you’re using alerting.
80 % of the alerts I’m getting are not because of some of my servers being fucked, but “DatasourceNoData” errors caused by the #Grafana instance itself being broken. There’s at least one of them every month. It’s ridiculous. Don’t use it!
Thought I'd set up a #grafana cloud free tier. That turned into a great reminder why you should filter your metrics. Both my home and cloud clusters send so many metrics, I ran out of metrics so quickly that I 15folded the limit. 😱
oh, I had gotten an idea for a great #raspberrypi project: A network status display.
For people with monthly data limits it's useful to know how much data has been used, so a little display showing stats from the gateway router would be a great idea. It could also show graphs and other cool things.
Obviously it is not really possible to add this functionality directly to the router and such routers don't really exist, so a raspberry pi would be needed to get information from the router via whatever APIs it has.
Of course, if the raspberry pi itself is the router, this is a lot more straightforward.
Or you could get a #bananapi R2, and have a gateway router with wifi, GPIO for LCD display, web server, and NAS storage all in one! Pretty epic!
@lamp OFC that board is a good option in lieu of bad #CPE|s and rarely good options for a lot of people...
That being said a lot of good #Router OSes allow logstashing and thus dashboarding with something like #Grafana...
But I do think that a reconfigureable status dashboard - whether it's a physical display or website in LAN [propably both would be done the same way] - would be cool...
Enlightening, isn't it? There are other empty blocks, but they are either fairly standard or are described elsewhere in the document.
If you are familiar with #helm, you won't despair because you have the power of analytics.enabled: false. That works on the rest of this chart and is the standard way to en/disable things.
It doesn't work that way.
Let me save you some time with the terrible new #github code search. Here is the actual syntax:
"analytics.reporting_enabled: false"
Parce que j'aime bien avoir la conso dans Home Assistant, mais je préfère la souplesse de #Grafana pour la visualisation de l'historique, et avoir la donnée disponible dans mes dashboards 🙂
Ça me fait toujours sourire de voir ce graphique de charge CPU de cette instance Mastodon.
Ça montre simplement que c'est un outil utilisé par des gens et qu'ils ont un rythme de vie absolument classique.
Les connaisseurs de #grafana : vous savez comment je peux changer les valeurs pour qu'elles s'affichent sous le format du nom du serveur et pas avoir le nom du champ avant ?
Exemple avoir accueil.maison ou http://accueil.maison
Ihr seht ein verwirrendes Diagramm mit der heutigen PV-Erzeugung und der jeweiligen Abweichung vom optimalen Einstrahlwinkel der jeweiligen Module (0° wäre optimal). 🤯
Alles in purem PL/pgSQL.
Ich denke, darüber muss ich mal einen sehr komplizierten Blog-Post schreiben ...
Spotted #Grafana on #BlueSky so my question is what do we have to do to encourage more companies to come to #Mastodon because Bluesky is very much still a beta