| GitHub repositories | https://github.com/appsinet |
| https://www.linkedin.com/company/appsinet |
Carles
- 7 Followers
- 158 Following
- 46 Posts
Web applications development. Services interconnections with API's or automation tools. Open source development whenever possible.
@ainmosni depending on the specific needs, you can also consider the old school but totally valid Zabbix. I'm still using it on my homelab, and not wanting to change.
Alerting works easy and well. But the dashboards are not as beautiful as on Grafana. I don't use them anyway
@niclas also, on the main vars folder we only have reusable variables that are needed on several roles. Like an email address or user account used in several services, for example. Specific variables for each role leave on the vars folder of that role. I think we got this folder structure after reading some best practices blog posts. Might be outdated, because we did this 2-3 years ago. But still works for us
@niclas we have a few playbooks on the root folder (4-5, not more). But these playbooks are basically a set of role executions with different tags. So, if we want to execute the full playbook, we run the command with the production or staging inventory, and that's all. If we want to run only a few things, we use the tags parameter to only execute those. All the logic is always inside a role, playbooks only orchestrate
@niclas we have this structure at work, and works well. We execute `ansible-playbook -i inventory/prod/hosts.yml playbook.yml`, and it gets the variables both on vars and vault
@ds I would like to self host my email for privacy concerns, but it's the last service I would recommend self hosting. Basically:
- when your computer is on maintenance, you can loose the emails sent to you
- you'll have a lot of bots trying to send email from your server
- You waste a lot of time configuring DNS MX, DKIM, SPIF, and so.. and anyway, some providers like Outlook just put your emails in the Spam folder
On a past job, we self hosted an SMTP service, and was a pain in the ass
@zef my personal recommendation: think about your backup strategy first, and then the tool that adapts better to that. It will always better to have a backup on the tool X, Y or Z than no backup, or not useful backup.
Apart from that, I heard a lot of cool things about Kopia. But probably I'm biased by my environment
@danielpetrica I'm going to do it sooner or later. Or, at least, it was the initial idea. I will take a look at what you said before doing anything. Anyway, I'm using Plausible for a few personal and friends websites, with not small traffic and any need of visitor analysis. Just take a look at visits and no much more. Plausible needs several containers, and looks a bit heavy for my use case. I read about Umami, and it looks more lightweight. In theory, I would need less resources
@danielpetrica yeah, the React CVE made a lot of damage.. didn't know about that breaking change, because I'm about to install it from zero. Sounds like good reasons to change, yes. Thanks for sharing!
@danielpetrica I cannot help with your question, but I'm curious.. why are you making that change? I'm about to make the inverse migration, because Umami looks more lightweight and has the features I need
@michabbb it is indeed a good tool. But in the development world, there's nothing like a silver bullet. Some advanced users would ask you why you are running docker compose, instead of any kind of kubernetes environment. And people working with clients that have a typical hosting would tell you that there's no way to use Docker there, but it's really easy to use deployer.
I'm using docker compose for my homelab, kubernetes on my job, and deployer for a friend's website.
I'm using docker compose for my homelab, kubernetes on my job, and deployer for a friend's website.