@d19dotca
that is also because Nginx uses FPM and not FastCGI(Apache in some cases can use FPM) for PHP, which give already a good boost.

NextCloud and WP could take good advantage of it, and Nginx could help solve the issue of the main image being too "fatty" 😉

It's easy to use Nginx in a multi-layer image.

@d19dotca This would be very helpful! +1

It seems they don't offer SMTP relay as per https://www.mailerlite.com/help/does-mailerlite-offer-smtp-simple-mail-transfer-protocol . Their partner company mailersend has it - https://www.mailersend.com/help/smtp-relay . If you can, please try using mailersend with Cloudron's External SMTP Server option - https://docs.cloudron.io/email/#relay-outbound-mails . If that works well, I can add it to the dropdown (the dropdown is really just a UI template to prefill things).

@d19dotca that's correct. There is no difference between using S3 compat vs Wasabi option since as you guessed they are just templates for pre-filling. Also, when you update, it won't get automatically changed to Wasabi, it will remain as s3 compatible.

The change is entirely just UI - https://git.cloudron.io/cloudron/dashboard/-/commit/8dfc142905171160ba41bdb592933bd2dd60748d

There is a much simpler way to keep things in sync in a p2p fashion with GunDB.

The trickier part is integrating with the different apps in a way where it can use that data.

@girish said in Mailing feature: add users and groups as members:

When added to a list, are you expecting the primary email of the user to be used ?

Indeed 🙂

This is now fixed for the next version. On upload conflict, the code will add .n to the filepath, whereas n is an increasing number until no conflict is found.

@girish I see, which seems like a fine choice for a single server, however QuestDB is better suited for multi server aggregation and immediate consistency from multiple sources.

@girish said in Include Size of App in Config>Storage & Backups:

Should have it in the graphs as well.

Updated the top post to include graphs ;-] 🙏

@wind-gmbh FWIW, we don't use the upstream distro packages. We use the packages straight from nginx.org since they provide better security fixes - https://nginx.org/packages/ubuntu/pool/nginx/n/nginx/ . Looks like https://nginx.org/packages/ubuntu/pool/nginx/n/ is the pre-built modules they have.

@andreasdueren said in Hide admin user(name)s for mail managers:

@girish personally I would enjoy the option to limit the E-Mail manager to the groups they are part of. For example some clients use so little resources like only a website with E-Mail that it doesn’t make sense to give them a dedicated server but instead put a few on the same.

Very good points as well, however as @girish says, as this would require to implement shared resources the idea is easier said than done. On the other hand, I personally can't see why it would not make sense to provide dedicated VPS at low cost for its own little resources needs to such client. In when you think about it, any client can grow and in such case the client is already set to scale without any more hassle. Cloudron is not too heavy for a simple basic use as it will adjust its own resources need only for the needs of the installed apps.

Now it would be nice if they could configure their own emails without being able to see let alone change the configuration of others.

Moreover, as explained previously if you install Cloudron for the client, a 10 bucks per month VPS would be more than perfect for such low needs, as superadmin you keep full control of the instance's management for the client and you give it admin privileges so it can fully manage its mail server. And with the upcoming multiple instances of Cloudron feature that should should become easier to manage from your side as well. 😊

Sounds like this is now done and live with 7.1?

@JLX89 this is currently not on our immediate roadmap.

@jdaviescoates Awesome, thanks -- I missed that.

Only workaround without adding the feature is to simply have Sufer or LAMP redirect to your dashboard. You should also be able to use an existing LAMP or Surfer app by checking the Host header.

@mehdi said in Email notify on SSH login:

@xarp actually, Cloudron uses the inbound port 25. This does not prevent it from sending stuff.

I tried the postfix/mailutils method which uses a simple bashrc method in the user/root home, however when Cloudron rebooted, the mail service couldn't start unfortunately.

Matomo is also not the only web analytics software in the Cloudron appstore, so a generalised solution would be better.

Since Cloudron manages the reverse proxy it could theoretically even inject the tracking code directly into the webserver response. An example of this can be found at https://stackoverflow.com/a/51664342.

@girish I believe they are separate, so need to be added separately as-such.

It seem that FileBase uses Storj as one of its layers, but adds additional features and redundancy beyond the Storj layer, hence the value-add higher pricing.

I just think both ethically aligned with the vision here, and having named support, as opposed to generic, is a nod to their efforts in providing a decentralised and competitive alternative to the incumbents.

I think recent events certainly show us that censorship is an ongoing concern, and replicated redundancy is a necessity to mitigate the risks of what might have been unlikely one day, and has suddenly happened another.

@nj not sure if this would be a good idea to map Cloudron dashboard labels of apps into the apps as such. It will be quite hard to explain where such labels within the app might come from. Further an app label is not required, so it may be unclear what to be used as the fallback default. Or maybe I haven't really understood what kind of label this might be. If you look for general env variable support, the cloudron cli tool support adding further env variables https://docs.cloudron.io/packaging/cheat-sheet/#environment-variables

@girish yeah, that's one option, as is the SSH tunnel, but thinking about a more user friendly approach so DBAs can simply do DBA things with a few clicks from the admin instead of setting up tunnels, etc.

@BrutalBirdie I think this somewhat depends on the use-cases. For production systems, a full restore is probably the best option to get back up. However for non-critical systems, often things are worth debugging to actually find the root cause and fix that to ensure the next update from the restored system does not break again.

@atridad are there any Drone feature requests to include a single runner by default and if one needs more, to add them separately?

@andreasdueren 7.1 is already out but as "unstable". You have to go to Settings -> Check For Updates and Click update button manually. Don't update unless you want to test things out already.

This bugged me again today.