Repository navigation
docs: update Cloud Registry page to 1.0.0rc1 - #92
Merged
Merged
Conversation
Bump the operator page of the ELIXIR On-Cloud registry from Cloud Registry v0.6.0 to v1.0.0rc1, and the front end from the branch ai-post-deploy to the Cloud Registry UI release v1.0.0-rc1. - Info box, overview, chart file (cloud-registry-1.0.0-rc1.tgz), images, verification commands and links now point to 1.0.0rc1; the "from 1.0.0rc1" warning and markers are gone. - New "Upgrading from 0.6.0", with the front end upgrade in the same step: verified release image pinned by digest, LOGIN_PROVIDERS instead of LS_LOGIN_URL/OIDC_*, the client ID in the audience, registry first. - The 0.4.0 and 0.5.0 paths are shortened to their deployment-specific points, with links to the product guides. - Replicas: cloudRegistry.replicas 2 with the chart's PodDisruptionBudget, instead of the warning. - Front end: environments and URL schemes come from the registry's service info, so NEXT_PUBLIC_SERVICE_* settings are dropped; the image build section is replaced by verifying the release image. Closes #91 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011FSoqMUX3EU35poH13xMgG
Contributor
Reviewer's GuideUpdates the Cloud Registry operator addendum from v0.6.0 to registry v1.0.0rc1 and UI v1.0.0-rc1, adding release-image verification, coordinated upgrade procedures, replica/PDB operations, and refreshed deployment configuration and links. Sequence diagram for the coordinated Cloud Registry upgradesequenceDiagram
participant Operator
participant Registry
participant UI
participant Users
Operator->>Registry: helm upgrade
Registry-->>Operator: rollout status
Operator->>Registry: kubectl rollout restart
Registry-->>Operator: rollout status
Operator->>UI: Deploy verified digest-pinned release image
UI->>Registry: Read service info and allowed settings
Users->>UI: Log in and use registry
UI->>Registry: Create, update, or delete services
Flow diagram for verified release image deploymentflowchart TD
Chart[Download release chart]
ChartVerify[gh attestation verify chart]
RegistryImage[Inspect registry image digest]
RegistryVerify[Verify registry image attestation]
UIImage[Inspect UI image digest]
UIVerify[Verify UI image attestation]
Deploy[Deploy verified chart and digest-pinned UI image]
Chart --> ChartVerify
RegistryImage --> RegistryVerify
UIImage --> UIVerify
ChartVerify --> Deploy
RegistryVerify --> Deploy
UIVerify --> Deploy
File-Level Changes
Possibly linked issues
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
Documentation build overview
|
uniqueg
marked this pull request as ready for review
November 2, 2026 10:09
Contributor
There was a problem hiding this comment.
Hey - I've found 5 issues
Prompt for AI Agents
Please address the comments from this code review:
## Individual Comments
### Comment 1
<location path="docs/operators/cloud-registry.md" line_range="1339" />
<code_context>
--authenticationDatabase admin --gzip --archive --drop' \
< cloud-registry-<timestamp>.archive.gz \
- && kubectl scale deploy/cloud-registry --replicas=1 -n <namespace>
+ && kubectl scale deploy/cloud-registry --replicas=2 -n <namespace>
```
</code_context>
<issue_to_address>
**Restore can mix live writes**
When a registry pod is still terminating or handling a request after scaling to zero, `kubectl scale --replicas=0` returns before the pods have necessarily exited, so `mongorestore --drop` can replace collections while a pod is writing; the restored data is then mixed with or overwritten by live writes.
Wait until all registry pods have terminated before running `mongorestore --drop`.
</issue_to_address>
### Comment 2
<location path="docs/operators/cloud-registry.md" line_range="118" />
<code_context>
+limit of services once the registry has them (see the UI's [upgrade
+notes][cr-ui-upgrading-1.0.0-rc1]). For this deployment:
+
+1. **Services per user:** the values file below limits users other than
+ admins to 20 services each (see [Services per
+ user](#services-per-user)), not to the default of 50. The app counts
</code_context>
<issue_to_address>
**Writes bypass upgrade safeguards**
When users or clients continue writing while Helm rolls out the new registry, old 0.6.0 pods continue serving requests during the rollout, so owners can register past the new 20-service limit or add organization fields after the pre-upgrade scan. Those writes persist when the old pods exit, leaving over-limit accounts and services that require cleanup; users then cannot register more services or update records that retain unsupported fields.
Pause writes for the upgrade and repeat the checks after all old pods have stopped, cleaning up any affected services before resuming writes.
Also at `docs/operators/cloud-registry.md:136`.
</issue_to_address>
### Comment 3
<location path="docs/operators/cloud-registry.md" line_range="51" />
<code_context>
| Admins | members of the LS Login group `ECP_CLN:ADMIN`, in admin mode (see [Admins](#admins)) |
-| Helm chart | `cloud-registry-0.6.0.tgz` of the [v0.6.0 release][cr-release] (deploys the image `elixircloud/cloud-registry:0.6.0`) |
+| Helm chart | `cloud-registry-1.0.0-rc1.tgz` of the [v1.0.0rc1 release][cr-release] (deploys the image `docker.io/elixircloud/cloud-registry:1.0.0rc1`) |
+| Front end image | `docker.io/elixircloud/cloud-registry-ui:1.0.0-rc1` of the [v1.0.0-rc1 release][cr-ui-release] (see [UI version](#ui-version)) |
| Database | MongoDB 7.0, deployed by the chart |
</code_context>
<issue_to_address>
**UI release instructions are unavailable**
When the page is used before the UI `v1.0.0-rc1` tag, image, and release documentation are published, the documented image cannot be pulled or verified, and the linked release pages are unavailable, so operators cannot complete the prescribed UI installation or upgrade.
Publish the UI tag, image, and linked release documentation before directing operators to use them, or document an available UI release instead.
Also at `docs/operators/cloud-registry.md:172`, `docs/operators/cloud-registry.md:850`, `docs/operators/cloud-registry.md:867-871`, `docs/operators/cloud-registry.md:1439-1440`.
</issue_to_address>
### Comment 4
<location path="docs/operators/cloud-registry.md" line_range="143" />
<code_context>
```bash
curl -fsS https://elixir-on.cloud/ga4gh/registry/v1/services \
- | jq '[.[].environment] | unique'
+ | jq '[.[] | select(.organization // {} | keys - ["name", "url"] | length > 0) | .id]'
```
</code_context>
<issue_to_address>
**Upgrade preflight misses hidden services**
When the deployment has pending or non-public services with organization fields other than `name` and `url`, when stored services are pending or otherwise non-public, anonymous `GET /services` omits them, so the jq preflight misses records with unsupported organization fields; after the upgrade their updates are rejected and owners were not warned beforehand. The startup warning occurs only after the change.
Inspect the full stored-service inventory with admin access before upgrading, or explicitly warn that hidden invalid services will only be identified by startup logs after the upgrade.
</issue_to_address>
### Comment 5
<location path="docs/operators/cloud-registry.md" line_range="1054-1055" />
<code_context>
+1. The service info shows version `1.0.0rc1` and the deployment's fields.
+ Its registration rules show the recommended settings: the environments
+ `["prod", "test"]`, the URL schemes `["https"]`, the service types
+ `controlled` (with `"max_unused_service_types": 0`, as only admins add
+ types) and `"max_services_per_user": 20`:
```bash
</code_context>
<issue_to_address>
**Service-info check reports a mismatch**
When operators use the prescribed values file without overriding `max_unused_service_types`, the values file leaves `max_unused_service_types` at its default of 10, while the documented service-info check expects 0, so the check reports a mismatch for a deployment using the supplied configuration.
Set `config.custom.access_control.max_unused_service_types` to 0 in the values file, or update the documented expected value to match the configuration.
</issue_to_address>Linked issues
#91 docs: update Cloud Registry page to 1.0.0rc1 · 5 of 5 met
| # | Requirement | Status | Notes |
|---|---|---|---|
| 1 | Update the page's version information, overview, verification and installation commands, and registry documentation links for v1.0.0rc1. | ✅ Met | The chart and registry image use the new versions, and the registry documentation references point to v1.0.0rc1. |
| 2 | Remove the “from 1.0.0rc1” warnings and markers. | ✅ Met | The warning and markers are removed, and the services-per-user setting is included in the values file. |
| 3 | Add upgrade guidance from 0.6.0, including the coordinated UI upgrade and its login and audience changes. | ✅ Met | The page documents the data-preserving registry upgrade, the verified UI release image, login variable changes, audience requirement, and registry-first order. |
| 4 | Use the chart's new cloudRegistry.replicas setting for the recommended two replicas. | ✅ Met | The values example sets cloudRegistry.replicas: 2 and the page describes the chart's PodDisruptionBudget. |
| 5 | Shorten older upgrade paths covered by product documentation while keeping links to those guides. | ✅ Met | The 0.4.0 and 0.5.0 paths are summarized with links to the registry guides and chart notes. |
Sourcery assessment
Approval pending. 5 findings to address first.
Blocking findings: docs/operators/cloud-registry.md:1339, docs/operators/cloud-registry.md:118, docs/operators/cloud-registry.md:51, docs/operators/cloud-registry.md:143, docs/operators/cloud-registry.md:1055
- Restore: wait until the app's pods are gone before mongorestore. - Upgrade from 0.6.0: pending services are not in the public list, the startup warning names them; writes through 0.6.0 pods bypass the new checks during the rollout, scaling to 0 first rules that out. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011FSoqMUX3EU35poH13xMgG
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Updates the operator page of the ELIXIR On-Cloud registry (
docs/operators/cloud-registry.md) from Cloud Registry v0.6.0 to v1.0.0rc1 (released 2026-11-02), and the front end from the branchai-post-deployto Cloud Registry UI v1.0.0-rc1, its first release. The page stays an addendum: it links the product docs and keeps only this deployment's values and decisions.cloud-registry-1.0.0-rc1.tgzof the v1.0.0rc1 release, imagedocker.io/elixircloud/cloud-registry:1.0.0rc1, new row for the UI image), verification and download commands, install command, checks (version1.0.0rc1,max_services_per_user: 20,max_unused_service_types: 0in controlled mode), and all registry links to/en/v1.0.0rc1/.max_services_per_user: 20is no longer commented out in the values file.curl | jqcheck before upgrading), the fully qualified default image, replicas, verify, upgrade and restart once. The UI upgrade is in the same step: release imagedocker.io/elixircloud/cloud-registry-ui:1.0.0-rc1, verified and pinned by digest;LOGIN_PROVIDERS/LOGIN_REDIRECT_URIinstead ofLS_LOGIN_URL/OIDC_*(the app stops if only the old ones are set, and warns if they are set next toLOGIN_PROVIDERS); the UI's LS Login client ID inconfig.security.auth.audience; registry first, then UI.cloudRegistry.replicas: 2in the values file, with the chart's PodDisruptionBudget, instead of the "chart can't set replicas" warning; the restore command scales back to 2; a new check lists the pods and the PDB.gh attestation verify); "Building the image" removed.NEXT_PUBLIC_SERVICE_ENVIRONMENTSis dropped from the settings, as the UI reads environments and URL schemes from the registry's service info (the "known limitation" of the create form is gone too). The Deployment example pinsdocker.io/elixircloud/cloud-registry-ui:1.0.0-rc1@sha256:<digest>.opt_in_roles: []workaround for front ends without the switch is only mentioned in the 0.5.0 path; all UI releases have the switch.Public page: no secrets, tokens or client IDs (placeholders only), no internal hostnames.
Sources: cloud-registry v1.0.0rc1 (release assets,
docs/releases/1.0.0rc1.md,docs/upgrading.md,docs/operations.md,deployment/README.md,cloud_registry/ga4gh/registry/service.pyfor the log lines), the RTD build at/en/v1.0.0rc1/, cloud-registry-uidevat 5a0a635 (docs/operator-guide/*,docs/user-guide/registering.md,docs/releases/1.0.0-rc1.md).Light lane: docs only, one page, no logic; reviewed the diff myself.
Verification
python scripts/check_names.py: passed.mkdocs build --strict: passed (anchor validation iswarninmkdocs.yml, so the strict build checks internal anchors); all#…links of the page resolve in the built HTML.pytest: 25 passed, 100% coverage./en/v1.0.0rc1/links exists on the fetched RTD pages (incl.upgrading/#from-060,#stored-services,deployment/#to-app-version-100rc1,#replicas,configuration/#services-per-user,access-control/#invitations).cloud-registry-1.0.0-rc1.tgzand.sigstore.jsonon v1.0.0rc1; the chart'sversionis1.0.0-rc1,appVersion1.0.0rc1.gh release download v1.0.0rc1 …,gh attestation verifyof the chart and ofoci://docker.io/elixircloud/cloud-registry:1.0.0rc1, and of the image by digest (@sha256:ea70b04a…, as the UI section does): all passed. Thejqcheck of organization fields was tested on sample data and against the live API (no hits).v1.0.0-rc1doesn't exist, so the UI links (cloud-registry-ui.readthedocs.io/en/v1.0.0-rc1/…, the GitHub release) and the UI image commands only work after the tag. Their anchors were checked against the UI'sdevsources and, where present, the currentlatestbuild.CI passed).mongorestore --drop; the 0.6.0 upgrade notes that the public list leaves out pending services (the startup warning names them) and that writes through 0.6.0 pods during the rollout bypass the new checks (scale to 0 first to rule that out).Declined review findings
v1.0.0-rc1, its image and its docs exist.max_unused_service_types: 0in the checks vs. the default 10: the service info of 1.0.0rc1 reports 0 in thecontrolledmode (release notes, #313/#316), so the check is right.Open questions
None.
Closes #91
🤖 Generated with Claude Code
https://claude.ai/code/session_011FSoqMUX3EU35poH13xMgG
Summary by Sourcery
Update the Cloud Registry operator page for the v1.0.0rc1 registry and UI releases, covering deployment configuration, verification, and upgrade procedures.
Enhancements:
Documentation: