Help
Paused and waking instances
Tested with:Databasezy API v1 · zb CLI 0.1
A paused instance has no compute running. Its storage and backups are kept. There are three kinds of pause, and they behave differently when a client connects:
- You or a teammate paused it. Connections are refused until someone resumes it. A connection never wakes it, so no compute is billed.
- Databasezy support paused it. Connections are refused, and only support can resume it.
- A Free-plan instance went to sleep because nobody used it. The first connection wakes it up.
A suspended instance is a separate case: it was stopped for billing or a quota, and resuming does not work until
that is resolved.
Why is my instance paused?
Section titled “Why is my instance paused?”- Someone paused it. From the portal,
zb instances pause, or the API. Pause and resume are recorded in the audit log with who did it. The instance’s status message sayspaused by its owner; connections are refused until it is resumed. - Support paused it. The status message says
paused by Databasezy support; contact support to resume it. - A Free-plan instance was idle. Free instances sleep after 15 minutes without connections and wake on the next one.
- A Free-plan instance was archived. After 7 days asleep, a Free instance is archived: its compute is removed, its disk is kept.
Paid instances never pause on their own.
paused: the instance is paused by its owner; resume it in the portal or with `zb instances resume`
Section titled “paused: the instance is paused by its owner; resume it in the portal or with `zb instances resume`”Cause: the instance was paused on purpose by a member of your organization. The gateway refuses the connection and
does not wake it. Retrying does not help: Postgres clients get SQLSTATE 08004, which drivers do not retry.
Fix: resume it from the instance page in the portal, or:
zb instances resume <instance_id>Resuming needs permission to operate the instance. It returns straight away; the instance is usable when its status
is ready again (zb instances get <instance_id>).
An instance that was paused before this behaviour shipped can answer the first connection after it with the instance is waking up; retry in a few seconds. It is not woken, and the following connections get the message above.
paused: the instance was paused by Databasezy support; contact support to resume it
Section titled “paused: the instance was paused by Databasezy support; contact support to resume it”Cause: support paused the instance, for example while an abuse report is reviewed. Connections are refused, and
a resume from the portal, the CLI or the API is refused with the same message (HTTP 409).
Fix: open a support ticket (see How to get help). Only support can resume it.
the instance is waking up; retry in a few seconds
Section titled “the instance is waking up; retry in a few seconds”Cause: a sleeping Free-plan instance is waking. The gateway holds the first connection for up to 30 seconds. If the instance is not ready by then, the client gets this error.
Fix: retry. Set your driver’s connect timeout to at least 30 seconds so the first connection after a sleep can wait for the wake. Opening the SQL console or another database tool in the portal also wakes a sleeping instance.
How long does waking take?
Section titled “How long does waking take?”Typically 10 to 40 seconds, depending on the engine and size. Paid instances do not sleep, so production workloads that cannot wait for a wake should run on a paid plan.
Archived Free-plan instances
Section titled “Archived Free-plan instances”An archived instance is shown as archived. Resume it from the portal or with zb instances resume to bring it back.
For how long an archived instance is kept, see Pause and resume.
ClickHouse cannot be paused
Section titled “ClickHouse cannot be paused”ClickHouse does not support pause. A pause request is refused with `clickhouse` does not support pause.
What is billed while paused
Section titled “What is billed while paused”Compute is not billed while an instance is paused. Storage and existing backups are kept and billed as usual. No new
scheduled backups are taken while an instance is paused, and a manual backup is refused with backups need a running instance. Take a manual backup before you pause if you need a recent one. See
Understanding your bill.
Suspended is not paused
Section titled “Suspended is not paused”These gateway errors mean the instance is suspended, not paused:
suspended: the instance is suspended (...); contact the account ownerquota_exceeded: the instance is suspended (...); usage resets next period or upgrade the plan
A resume request on a suspended instance is refused with instance is suspended for billing reasons. Fix the
billing problem or the quota first. See Spending caps and quota errors
and Invoices and payment. An instance you had paused before it was suspended
goes back to paused when the suspension is lifted; it does not start running.
Keep a staging database off overnight
Section titled “Keep a staging database off overnight”Run zb instances pause <instance_id> and zb instances resume <instance_id> from a scheduler (cron or a CI
schedule), or call POST /v1/orgs/{org}/instances/{id}/actions/pause and .../actions/resume. Resuming is not
instant, so schedule it a few minutes before people need the database.
Still stuck?
Section titled “Still stuck?”Search the help center or ask the assistant in the portal Help panel. If that does not solve it, open a support ticket from the portal (Help → Open a ticket). Include the instance id, the region, its status and plan, the exact error text, your client or driver version, and the time it happened in UTC. Never paste passwords, connection strings with passwords, API keys or patient data (PHI) into a ticket. See How to get help.