Instances overview
An instance is one running deployment of a technology — a PostgreSQL cluster, a MongoDB replica set, a Valkey cache. Whatever the technology, you find, inspect and manage instances in the same place.
The Instances page
Instances in the left menu lists every instance you have access to, across all namespaces. For each instance you see its Status, Name, Provider and Namespace, and an Actions menu.
- Use the search, filter and column icons above the table to narrow the list.
- Create opens a panel where you pick the provider for a new instance. See Create an instance.
- Import database creates an instance from a backup that already exists in a backup storage. See Create an instance from a backup.

When no instance exists yet, the page shows the installed providers instead, so you can start right away.
Instance status
The status column shows one status word. It is the same word in the UI, in everestctl and in kubectl get instances. Common ones:
| Status | Meaning |
|---|---|
| Creating | The instance has never been up yet. The provider is creating it. |
| Ready | The latest configuration is applied and every check passes. |
| Applying changes | A change, such as scaling or a configuration change, is being rolled out. |
| Upgrading, Restoring, Resizing storage | A long operation is running. |
| Pods starting | Pods are not ready yet. |
| Pods can't be scheduled, Crash loop, Image can't be pulled | Pods have a problem. |
| Invalid configuration | The provider rejected the instance's settings. The message names each invalid field. |
| Secret not found, Backup storage not found | An object the instance uses is missing. |
| Provider not found, Provider not responding | No provider is working on the instance. See When no provider reconciles the instance. |
| Deleting | The instance is being deleted. |
The colour tells you whether to wait or act. Yellow means work is in progress, or a problem that may still clear by itself. Red means progress has stopped and someone has to act: either at once, for example after an invalid configuration, or after a problem has lasted 10 minutes.
While an instance that has been up before is changing, the status also says whether clients can connect, for example Upgrading · available or Restoring · unavailable.
When something needs attention, the instance page shows an alert with the reason and what to do, for example "engine: 1 of 3 pods cannot be scheduled: 0/2 nodes are available: 2 node(s) didn't match pod anti-affinity rules." While a restore, an upgrade or a storage resize runs, the instance can't be edited. See Diagnostics.
Instance details
Click an instance to open its details. The page has these tabs:
- Overview — basic information, connection details, and cards with the instance's configuration, grouped as the provider defines them (for example Resources, Proxy or Advanced configuration). Cards that can be changed have an Edit button. See Change and delete instances.
- Backups — the instance's backups and backup schedules, and point-in-time recovery. See Create backups.
- Restores — the history and status of restores into this instance. See Restore a backup.
Generic plugins can add their own tabs and overview cards, for example to browse data or show metrics.
Instances in Kubernetes
Every instance is an Instance custom resource. You can see the same information with kubectl:
kubectl get instances -A
NAMESPACE NAME PROVIDER VERSION AVAILABLE STATUS AGE
demo catalog-db percona-server-mongodb 8.0.32 False Creating 14s
demo orders-db provider-percona-postgresql 18.4-1 True ApplyingChanges 26m
demo session-cache valkey 9.1 True Ready 3d
STATUS is the reason of the Ready condition. kubectl get shows the reason as one word, for example ApplyingChanges, where the UI shows Applying changes. Add -o wide for the phase, whether the instance is stalled, and the message.
To wait until an instance is ready:
kubectl wait instance/orders-db -n demo --for=condition=Ready --timeout=10m
For the full status, including every condition and the number of ready pods per component:
kubectl get instance orders-db -n demo -o yaml
See Instance.