Cloudron Forum

The 500 status code from your screenshot also hints at a server side exception or crash, maybe worth looking into the app server side logs while that happens. Maybe the crash can then be reported upstream for that faster search module

Looks like Cloudflare is interfering here and changing the data in transit? You would have to contact Cloudflare about this, but I have no high hopes that this would get anywhere.

@vadim said in Running a separate PostgreSQL instance alongside Cloudron: Is there any way to use pgvector via any cloudron app? The answer is a simple search for "pgvector" in the forum. It yields: https://forum.cloudron.io/topic/13139/consider-adding-support-for-pgvector-pgvectorscale-for-pgsql

@james seems to me this is the likely cause of this too: https://forum.cloudron.io/topic/13068/issue-with-email-delivery-going-through-cloudron-mailing-lists

The problem is only when email is forwarded by Cloudron. If it is sent directly to Gmail there is no any problem.

Ok! Thank you! Cloudron in any case amazing product

@vadim thank you very much

Fixed! Thank You!

@vadim OK, so, the root cause is the DNS resolution issue . Looks like it starts to fail at some point. Is there something in the network that blocks DNS requests? Looks like this is some internal network? Maybe try to set up a forward zone to the internal DNS (https://docs.cloudron.io/networking/#internal-dns-server) ? In the next release, we are moving away from unbound since it's causing not-so-easily debuggable situations (like this one).

@vadim What I means is that you have to start from the latest elastic skin and then make whatever changes you made previously on top of the latest skin. Is your skin heavily customized? If so, it might be harder. Unfortunately, I have no experience how to manage roundcube skin changes.

I got it, thx. I will check

@vadim that looks like a public key and not the private key.

Ah ok. @vadim we use freescout as our primary support tool , can say it is very good and stable (been using it over 6 years now).

This was sorted out in support.