Troubleshooting
The three failures that account for most support requests, each in the order worth checking. If none of it helps, the end of each section says exactly what to send support so the first reply is an answer rather than a question.
Email is not working
Mail is not arriving
- Check Junk. Legitimate mail is filtered more often than people expect.
- Check the quota. A full mailbox rejects incoming mail and the sender gets a bounce. Portal → Emails → the account. See quotas.
- Check the MX record. Portal → Domains → the domain. Without a correct MX record, sending servers have nowhere to deliver. A recent DNS change can take 30 minutes to be visible — see propagation.
- Check the address. Ask the sender to read back exactly what they typed.
Mail is not going out
Check sending is configured first. Outgoing mail is handed to an SMTP relay, and until one is set up under Server → Email → Sending, mail is held in the outbox rather than delivered. It is not lost — it goes out when a relay is configured. This is the single most common cause and it looks exactly like a broken mail server. Full explanation.
After that:
- Client settings. Port 587 with STARTTLS. Not port 25.
- Password. Sign into webmail to test it independently of the client. If webmail works and the client does not, the client has a stale password.
- Account status. A suspended account cannot send.
Mail arrives, but in the recipient’s spam folder
Almost always incomplete authentication on your domain:
- SPF missing — the receiving server cannot confirm your relay is allowed to send as you.
- DKIM missing — nothing signs the message.
- DMARC missing — no policy telling receivers what to do about the first two.
- A young domain. New domains have no sending reputation. It builds over weeks of consistent legitimate sending; nothing makes it instant.
SPF, DKIM and DMARC explained covers all three properly. Support can check your current records if you would rather not.
Webmail will not load
Confirm the full URL — https://mail.yourdomain.com/webmail/, including the path. Try
another browser to rule out a cached asset. If it is “site can't be reached” rather than
an error page, the request is not arriving at all and the service needs restarting; contact support.
What to send support
The address, whether it is sending or receiving that fails, the exact error text, and when it started.
An application will not start
Read the status first
| Status | Meaning |
|---|---|
| Running | Healthy. |
| Starting | Coming up. Give it 30–60 seconds. |
| Stopped | The container exited. Restart it. |
| Error | It failed to start. The logs will say why. |
Restart once. If it comes straight back to Error, restarting again will not help — read the logs.
The logs
Apps → the app → Logs. The message is usually explicit:
| Log line | Cause |
|---|---|
Error establishing a database connection | Wrong credentials, or the database container is down. |
Permission denied | File ownership inside the container. |
Address already in use | Port conflict with another app. |
Out of memory / killed | The server does not have the RAM for this workload. |
WordPress: database connection error
The most common WordPress failure, and it has two causes:
- The database container is down. Servers → the server → Databases — check MySQL is Running.
- The credentials drifted. Changing the database password without updating
wp-config.phpleaves WordPress authenticating with the old one. Support can re-sync them.
It broke right after a change
Then the change is the cause, and reverting it is the fastest diagnosis. For WordPress, the blunt version of that is deactivating every plugin at once over SSH:
mv wp-content/plugins wp-content/plugins.bak If the site comes back, one of them is responsible. Restore the directory and re-enable them one at a time.
It sits on “Starting” forever
More than two minutes means the container is waiting on something it cannot reach — usually a database or a volume. Restart once; if it repeats, contact support rather than looping.
What to send support
The app name and server, the last 20 log lines, what you changed before it broke, and whether restarting did anything.
A domain is not resolving
Work down this list in order — each step rules out the one below it.
1. Is there an A record?
Domains → the domain. There must be an A record pointing at the server's IP, plus
one for www if you use it. If there is not, that is the whole problem —
how to add one.
2. Has it propagated?
Look up the domain's A record with any DNS lookup tool. If the IP that comes back is your server's, DNS is correct and the problem is further down this list. Typical propagation is 5–30 minutes; a nameserver change rather than a record change takes up to 24–48 hours.
3. Is your own machine caching the old answer?
Entirely possible even after propagation completes everywhere else.
# Windows
ipconfig /flushdns
# macOS
sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder Then try it in a new tab, not a refreshed one.
4. Is the domain verified?
An Unverified domain is not yet linked to anything on your account. Click Verify Domain — see verifying ownership.
5. Is the app running?
A domain can resolve perfectly and still show an error, because resolving only gets the request to the server. Apps → the app should read Running. If it does not, start there.
6. Is the domain actually attached to the app?
Apps → the app → Settings. The domain must be selected there, or nothing routes the hostname to the container — see connecting a domain.
What to send support
The domain, the server's IP, a screenshot of your DNS records, and when the change was made.