Skip to content
OpenEverest Documentation

Get enterprise-grade support and services for OpenEverest from certified partners, or join community.

Get support

What’s new in OpenEverest 2.0.0 Developer Preview 4

Developer Preview — not for production

Not feature-complete; for testing and feedback only.

Install on a fresh cluster. There is no upgrade path from Developer Preview 3, and none from v1 yet. Running v1 and v2 side by side in one cluster is not supported.

A v1 → v2 migration path is planned and ships with v2 General Availability. It cannot exist during the preview because the v2 API is still changing and a migration tool needs a stable target. Keep running v1: it stays supported, and you will not be asked to rebuild what you are running by hand.

Breaking changes to the Instance and Provider CRDs, the provider-runtime SDK and RBAC policies — see Breaking changes. Nothing between previews gets a migration path, deliberately, so the API can be corrected before GA.

New to OpenEverest? Start with the Quickstart Guide, or read the blog post on the v2 architecture.

Release highlights

Instances say why their pods are not running

The runtime now fills status.components from each component’s pods (replicas, readyReplicas) and sets two conditions: PodsScheduled turns False once a pod has waited more than a minute for a node, and PodsReady names the worst problem (CrashLoopBackOff, ImagePullBackOff, CreateContainerConfigError, NotReady). Messages have one line per component, 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.”

The UI shows a warning banner and icon on affected instances, and everestctl instance status gains a READY column. It works for providers that label their pods with Context.PodLabels().

Pod placement in the UI, with defaults that hold

The creation wizard and the instance overview can now edit each component’s affinity rules. A provider enables the editor from its UI schema with widgetType: podSchedulingPolicy.

The defaults are now explicit: omitting affinity or topologySpreadConstraints gets the provider’s default, and an empty value ({} / []) means none. The MongoDB and MySQL providers now place each data-bearing member on its own node, so a three-member cluster needs three schedulable nodes unless you set affinity: {}. A cluster that is too small shows up as PodsScheduled=False instead of quietly doubling up on one node.

Optional sections in instance forms

Provider UI schemas can put optional settings behind an Enable switch with groupType: toggleable. While a section is off, its fields are hidden, not validated and not sent. A section starts on when the instance, preset or source backup already sets one of its fields; otherwise it starts off and the overview shows it as Disabled. Turning a section off while editing an instance removes its saved values. groupType: bordered draws the same card without the switch.

Time-based backup retention

Schedules can now keep backups for a period instead of a count: retention: {type: time, duration: 30d} (Nd, Nw or Nm). The PostgreSQL provider supports it. The MongoDB and MySQL operators only support count retention, so those providers report type: time on the BackupConfigured condition. The UI still edits count retention only.

Plugins that look like the core

Plugins can now match the core’s look without depending on its internals. The host shares only React (through an import map) and its design tokens (--everest-* CSS variables). Each plugin bundles its own MUI, and the new @openeverest/plugin-theme package themes it from those tokens, including live dark mode: wrap the plugin UI in <PluginThemeProvider cacheKey="my-plugin" nonce={api.cssNonce}>. A plugin is now skipped if it fails spec.compatibleHostVersions (not enforced before) or the new spec.compatibleUiContractVersions.

⚠️⚠️⚠️ Breaking changes for provider developers and administrators

1. schedules[].retentionCopies → schedules[].retention on Instance and InstancePreset: retentionCopies: 7 becomes retention: {type: count, count: 7}. After the CRD upgrade retentionCopies is pruned, so those schedules keep every backup until retention is set.

2. Context.Apply uses server-side apply (field manager provider-<name>, forced ownership) instead of Get + Update. Build the full desired object every time: fields you stop setting are removed, and fields you never set are left to the operator. Objects read with Get are rejected, and obj is no longer updated with the server’s response. Empty non-pointer omitempty structs are dropped instead of sent as {}.

3. status.components is computed by the runtime. ComponentStatus is now name/selector/replicas/readyReplicas (was podRefs/total/ready/state), and controller.Status.Components and controller.ComponentStatus are removed. Call c.PodLabels(component) in Sync and put the labels on that component’s pods; the provider needs list and watch on pods.

4. ComponentSpec.name is removed; the spec.components map key is the name. GetComponentsOfType() and Context.ComponentsOfType() return map[string]ComponentSpec.

5. default: true → defaultVersion. Version bundles and component versions no longer carry a default flag. Set defaultVersion at the top of the Provider spec and on each component type instead; the provider-sdk’s generate emits it from definition/versions.yaml (openeverest/provider-sdk#47). Without it, an Instance that omits spec.version gets no version bundle.

6. SchedulingPolicy.topologySpreadConstraints is now a pointer (*[]corev1.TopologySpreadConstraint), so unset and [] mean different things. Build a component’s constraints with controller.TopologySpreadConstraints(policy, c.PodLabels(component)).

7. RBAC. backups and restores grants now match the owning instance (<cluster>/<namespace>/<instance>) instead of the backup’s or restore’s own name. Reading instance credentials (.../instances/{instance}/connection) needs the new read-connection action; read no longer covers it. Policies with unknown action names are rejected. Known issue: read-only users see a 403 error on the instance page until the UI checks read-connection.

Changes

Added

  • Pod status and conditions on instances, shown in the UI and everestctl (#3252 by @shivansh-source, #3335, #3312, #3338).
  • Pod scheduling policy editor in the UI (#3294).
  • Toggleable form sections and bordered groups in the ui-generator (#3278).
  • BackupImport API and provider-runtime support, with no UI yet (#3126, #3138). (by @chilagrow)
  • Time-based backup retention (#3222 by @AdityaPimpalkar).
  • controller.TopologySpreadConstraints and documented scheduling defaults (#3309, #3311); top-level defaultVersion (#3264); read-connection RBAC action (#3288 by @VijetaPriya47).
  • @openeverest/plugin-theme 0.1.0 (#3076) and @openeverest/plugin-sdk 0.4.0 (#3283).
  • Nightly release e2e tests for PITR and restore to a new cluster (#2925, #3261).
  • Locci Cloud in ADOPTERS.md (#3243 by @MikeTeddyOmondi).

Changed & Improved

  • Breaking changes, detailed in Breaking changes (#3222, #3240, #3282, #3335, #3337, #3264, #3309, #3204, #3288).
  • provider-runtime no longer depends on any github.com/percona/* module (#3285).
  • Go 1.27.1. A provider that bumps core needs a Go 1.27 toolchain to build, so pinned toolchains such as golang:1.26 images or golangci-lint older than v2.13.0 must move too (#3280).
  • CI moved to CNCF-hosted runners and kind, into a single pipeline with merge-queue support, and from MinIO to SeaweedFS (#3205, #3210, #3211, #3213, #3216, #3217, #3218, #3249, #3259).
  • Contributor docs fixes (#3199 by @bhuvan-somisetty, #3202 by @Sarthak-Shreshtha01).

Fixed

  • Security: the restore endpoints enforced no RBAC (#3204 by @VijetaPriya47); read-only roles could read instance credentials (#3288 by @VijetaPriya47).
  • OIDC requests had no response size limit or timeout (#3219 by @alishair7071); an unresponsive PMM endpoint could hang monitoring reconciliation (#3205).
  • A seeded Instance went back to Restoring after its source Backup was deleted (#3257).
  • UI: switching topology kept the old topology’s fields (#3255); overview cards jumped between columns (#3316); fields without validation were treated as required (#3294).
  • everestctl: upgrade failed on the everest-crds release after the first upgrade (#3241); namespaces list ignored --json (#3248 by @jhalak101205-cpu).
  • CI fixes for fork PRs, go-consistent, /assign on edited comments and a flaky e2e step (#3245, #3258, #3244, #3308).

Installing OpenEverest 2.0.0 Developer Preview 4

v2 is installed with Helm; you need a running Kubernetes cluster and helm.

helm repo add openeverest https://openeverest.github.io/helm-charts/
helm repo update
helm install everest-core openeverest/openeverest \
  --devel --version "2.0.0-dev.4" \
  --namespace everest-system --create-namespace

# admin password
kubectl get secret everest-accounts -n everest-system \
  -o jsonpath='{.data.users\.yaml}' | base64 --decode | yq '.admin.passwordHash'

# UI
kubectl port-forward svc/everest 8080:8080 -n everest-system

everestctl is not published as a binary with this release: its install/uninstall commands still target v1 and cannot install v2, so shipping it would be misleading. To use the day-2 commands against an existing v2 installation, build it from source with make build-cli on the v2.0.0-dev.4 tag.

Providers are released independently. Install the ones you need from their own repositories, picking a version built against 2.0.0-dev.4 — earlier provider releases target the previous Instance and Provider schemas and will not work correctly.

v1 support and end-of-life timeline

Stage Description
Now v2 Developer Preview — testing and feedback. v1 remains the released, supported version.
v2 GA Expected in a few months, together with the supported v1 → v2 migration path.
GA + 3 months v1 enters Maintenance Mode: security patches and critical fixes only.
GA + 12 months v1 reaches End of Life.

You do not need to plan a migration during the preview, and you will not be doing one by hand.

Get involved

Build a provider against the new API and tell us where it hurts. Provider plugins, generic plugins and core improvements are all welcome.

Thanks to our contributors

@AdityaPimpalkar, @alishair7071, @bhuvan-somisetty, @jhalak101205-cpu, @MikeTeddyOmondi, @Sarthak-Shreshtha01, @shivansh-source, @VijetaPriya47

Full Changelog: https://github.com/openeverest/openeverest/compare/v2.0.0-dev.3…v2.0.0-dev.4