FAQ¶
Getting started¶
How do I create a Verda account and add funds?¶
Visit https://console.verda.com/signup, fill out the registration form, verify the account email, and log in to the console. From there, go to Billing → Top-up to add funds: add a card directly on that page, or set up bank transfer under My Account → Bank transfer. See pricing and billing for every top-up method, including SEPA and bank invoice.
How do I add a payment method and avoid running out of balance?¶
Add a payment method from Billing → Top-up (add a card there directly, or set up bank transfer under My Account → Bank transfer), then turn on automatic top-up from the same page so new requests aren't blocked the moment the balance runs out. Set the top-up trigger threshold high enough to cover at least roughly 12 hours of usage, so a slow bank transfer or a single declined card doesn't take resources offline. See pricing and billing for how automatic top-up and the low-balance threshold work.
How do I invite teammates, and what can each role do?¶
Invite a teammate from Team in the console: the Invite button opens an invite dialog, where you choose a role of Admin or Developer (the project's Owner is whoever created it, not an invite option). Roles determine what they can do: Owner has billing and payment-method access, Admin can view billing but not edit it, and Developer can use resources but has no billing access at all. See team projects for the full permissions table and how to change someone's role.
Should I use on-demand or spot for my first deployment?¶
On-demand gives dedicated capacity that stays allocated until the instance is deleted, and takes priority when capacity is tight. Spot runs on spare capacity at a lower rate but can be reclaimed at any time without warning. Use on-demand for anything that can't be interrupted, and spot only for interruptible or checkpointed work. Both are chosen when you create an instance.
Where can I check current GPU pricing?¶
On-demand and spot rates for every GPU are listed on the pricing page and in the console when creating an instance. Rates vary by model, configuration, and region, so the console is the reliable source at the time of deployment.
Account & billing¶
Why am I still charged after I shut down my instance?¶
Shutting down an instance powers off the operating system, but the GPU stays reserved for that instance, so billing continues at the full hourly rate. Any attached storage keeps billing too. Shutdown is meant for maintenance tasks (attaching, detaching, resizing, or cloning a volume), not for pausing costs. To stop compute charges, delete the instance instead. See shutdown and delete.
How do I stop all charges but keep my data?¶
Delete the instance, but leave its volumes unselected in the delete dialog so they're kept (nothing is selected by default). Compute charges stop immediately, and the data stays on the detached volume, which can be reattached to a new instance later. A kept volume keeps billing on its own until it's deleted too. See shutdown and delete for what the delete dialog does with attached storage.
My balance is dropping even though nothing is running. Why?¶
The most common cause is storage that's still billing on its own: a block volume keeps billing for as long as it exists, whether or not it's attached to a running instance, and deleting an instance does not delete its volumes. A separately provisioned shared filesystem also bills independently. Check Storage → Block volumes (and shared filesystems) in the console for anything no longer needed. Turning on Remind when storage is unattached under Billing → Billing notifications flags this automatically going forward. If the charges still don't add up, Support can pull a usage report. See deleting storage.
Why was my payment declined?¶
Most declines come from the card issuer or an unfinished bank-verification step. The decline code shown on the billing page indicates which:
| Decline code | What it usually means |
|---|---|
payment_intent_authentication_failure |
The 3D Secure / OTP step wasn't completed. Finish the bank prompt and retry. |
card_velocity_exceeded |
The card hit a transaction or spending limit. Contact the card issuer or use another card. |
stop_payment_order / revocation_of_all_authorizations |
The bank was told to stop the charge. Ask the bank to reverse it, or use a different card. |
generic_decline / do_not_honor / insufficient_funds |
The bank declined it directly; contact the bank or try another card. |
processing_error |
A temporary glitch; retry. |
Repeated rapid attempts across several cards can also trigger a temporary account-level security block; if that happens, wait 30 to 60 minutes before trying again. It's also worth confirming the billing details on file match what the bank has on record, and that the card is enabled for online or international payments.
How do I get an invoice?¶
A receipt is available after any payment, under Billing → Billing history. A proper invoice has to be requested from Sales before paying (available for top-ups at or above roughly 1,000 EUR/USD).
What can Owners, Admins, and Developers each do?¶
Only the project Owner can manage billing: top up, edit auto top-up and payment cards, view billing, transfer funds, manage team members, and rename or delete the project. Admins can view billing but can't edit it. Developers can deploy instances, create volumes, and manage resources, but have no billing access at all. See team projects for the full permissions table.
Technical troubleshooting¶
I can't SSH into my instance (permission denied, connection refused, or a timeout)¶
Work through these in order:
- Confirm the instance shows as running in the console, and use its current public IP (this can change after a rebuild).
- Ping the instance to confirm it's reachable at the network level. If ping fails, the instance may be down or a firewall rule may be blocking it.
- Check that port 22 is open, for example with
telnet <instance-ip> 22. A timeout or refusal usually means the SSH service isn't running or a firewall is blocking the port. - Try connecting from a different network, to rule out a local firewall or ISP-level block on port 22.
- Connect with verbose output to see exactly where the connection fails:
ssh -v -i /path/to/private-key <username>@<instance-ip>. Common causes at this point are the wrong username (tryroot,ubuntu, or the OS default), the wrong IP, or the wrong key. - Confirm the private key has restrictive permissions:
chmod 600 /path/to/private-key. SSH refuses a key that's too open (for example644or777). - If none of the above resolves it, generate a new key pair and redeploy the instance with it.
- If it still fails after that, contact Support with the instance ID, the exact command used, and the verbose SSH output.
Note
If you see a warning that the host key or fingerprint has changed: that's expected after a rebuild, or after a new spot instance reuses an old IP: clear the old entry with ssh-keygen -R <instance-ip> and reconnect. If the warning appears on an instance that wasn't rebuilt, confirm the IP being used is correct before continuing.
Once connected, a normal SSH session should open without further prompts. See troubleshooting SSH connection issues.
I lost my SSH key. How do I get back in?¶
If still connected to the instance through another session, add the new public key to ~/.ssh/authorized_keys (or /root/.ssh/authorized_keys). If locked out entirely: shut the instance down, detach its boot volume, attach that volume to a new instance as Additional storage, mount it, and add the new key to the authorized_keys of the user you log in as. Then unmount it, detach it, and reattach it to the original instance in place of its boot volume. See recovering lost SSH access for the full walkthrough.
My instance was deleted because my balance ran out. Can I get my data back?¶
The instance itself can't be restored, but its data volume is kept in Storage → Block volumes → Deleted storage for 96 hours after deletion. Within that window, restore the volume and, when creating a new instance, choose Use existing under Main OS volume to select it. After 96 hours the data is permanently removed. Turning on automatic top-up and the Warn about low balance alert (both under Billing → Billing notifications) avoids this happening again. See deleting storage for the restore window and what a restore costs.
My Serverless Container deployment is stuck at 0 replicas. Is that a problem?¶
Not necessarily. If the deployment's minimum replicas is set to 0 and it's idle, scaling to zero is expected, and a new request triggers a cold start back up. Keeping a minimum of 1 replica avoids the cold start but costs more, since that replica is always billed. If minimum replicas is set above 0, or the deployment is receiving traffic but still shows 0 replicas, check its logs, health checks, and scaling configuration; contact Support if several deployments are affected. See scaling and health checks.
I'm getting No space left on device. What should I check?¶
First identify which filesystem is actually full, for example with df -h. Remove unnecessary temporary files and caches, or move persistent data onto an attached shared filesystem or the deployment's persistent scratch disk where available. For storage-quota errors specifically: deleted volumes sit in Deleted storage for 96 hours and still count against quota during that window, so permanently deleting ones no longer needed there can free it up immediately. If more is genuinely needed, request a quota increase under My Account → Quotas. See quotas for how to request an increase and track it.