another week has flown by and while typing this introduction I am sitting in our Hanover office since my wife and I are joining her companies christmas party tomorrow afternoon. That is one of the benefits of a mostly web-based workflow, you can work from almost any device.
Startpages Link to heading
Speaking of web-based worklow.. for organising ones own webservices so called “startpages” continue to be a trend. Afterall they are the portal to all the little urls and certainly contribute to a higher WAF. Especially two names can often be found when talking about these pages Organizr and heimdall, but for my test both are a bit too heavy and feature rich, which is why I have so far settled for a custom fork of minimum-viable-startpage. Recently I have come across Homer which also seems interesting. Especially since its using yaml for configuration and it even is implemented as a PWA, which means its service worker can even serve pages when your webserver is currently offline.
Franz & Ferdi Link to heading
Another form of a start page are portal like applications such as Franz and its fork Ferdi. In a way both Franz and Ferdi are similar to what we wanted to provide with the Kopano DeskApp at my place of work, only that it supports more webservices and while the already mentioned Progressive Web Apps (PWA) can already make web applications feels more at home on the desktop, sometimes one may want to go a different route and “bundle” all the often used web applications in a single window. Ferdi emerged when the Franz developers forced account sign in on their software, and it not only removes the need for an account, but also gives the ability to run your own account server to centrally manage applications.
In case your own application is not yet available as a service, then there are extensive docs and example plugins on how to write your own integration.
Kubernetes as a Service with Rancher2 at Hetzner using Terraform and Helm Link to heading
I ended the previous issue with a mention of ManageIQ in hos blog Evaggelos shows an approach that is more in the spirit of personal self-hosting. He is using Terraform to provision machines at Hetzner to which he then deploys Rancher for Kubernetes and lastly uses Helm to deploy a first example service to the new cluster.
Add health checks to your services! Link to heading
No, I am not referring to health checks for containers, but the program/service called Healthchecks.io. For some uses cases a full blown (uptime) monitoring solution may be overkill and this is where services such as Healthchecks.io come in handy.
The service basically offers http endpoints that you can query from e.g. cron. If a request is getting behind (e.g. backup cron did not complete in time or at all) Healthchecks.io can send you a notification about it. Honestly its one of the few services I still run externally, since “what good is a monitoring solution if its down with the rest of your infrastructure?” :-D
And it would not be in this newsletter, if you would not be able to run it on your own infrastructure: https://github.com/healthchecks/healthchecks
Do you know of a project that should be featured here? Link to heading
I am always looking for new projects to try out! Just send a mail to selfhosted-newsletter@9wd.eu.
I hope you have enjoyed this issue. If this newsletter was useful to you please recommend it to colleges or tell them to subscribe to it via RSS.
Have a nice weekend! Felix