Monitoring API: remove maximum retention days prometheus

[marioferh]

Remove Maximum validation and documentation for durationInDays (365)
and sizeInGiB (16384) on Prometheus Retention. We do not have enough
customer or operational data to pick defensible API-level ceilings;
arbitrary caps could block valid large-cluster configurations or
misalign with what CMO/Prometheus can actually support.

[openshift-merge-bot]

Pipeline controller notification
This repo is configured to use the pipeline controller. Second-stage tests will be triggered either automatically or after lgtm label is added, depending on the repository configuration. The pipeline controller will automatically detect which contexts are required and will utilize /test Prow commands to trigger the second stage.

For optional jobs, comment /test ? to see a list of all defined jobs. To trigger manually all jobs from second stage use /pipeline required command.

This repository is configured in: LGTM mode

[openshift-ci]

Hello @marioferh! Some important instructions when contributing to openshift/api:
API design plays an important part in the user experience of OpenShift and as such API PRs are subject to a high level of scrutiny to ensure they follow our best practices. If you haven't already done so, please review the OpenShift API Conventions and ensure that your proposed changes are compliant. Following these conventions will help expedite the api review process for your PR.

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

📝 Walkthrough

Walkthrough

The PR converts Prometheus retention to an hours-based field in the API: the Go type Renamed DurationInDays → DurationInHours (int32) while preserving the JSON name durationInDays for wire compatibility, updates the PrometheusConfig.Retention comment to state an omitted default of 360 hours, and sets a minimum of 1 hour (the prior days-based maximum was removed). The generated ClusterMonitoring CRD schema updates the retention field’s documentation/validation to hours and increases retention.sizeInGiB maximum from 16384 to 2147483647.

Suggested reviewers

• JoelSpeed

🚥 Pre-merge checks | ✅ 15

✅ Passed checks (15 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title describes the main change: removing maximum retention validation for Prometheus, though 'days' phrasing is slightly imprecise given the broader scope.
Description check	✅ Passed	The description clearly explains the rationale for removing maximum validation constraints on both durationInDays and sizeInGiB, directly addressing the PR's core intent.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Stable And Deterministic Test Names	✅ Passed	All 137 Ginkgo test names are stable, descriptive, and static with no dynamic content like generated IDs or timestamps.
Test Structure And Quality	✅ Passed	PR contains no Ginkgo test code—only API type definitions and CRD schema changes. Custom check for test structure is not applicable.
Microshift Test Compatibility	✅ Passed	This PR does not add any new Ginkgo e2e tests; it only modifies API type definitions and CRD manifests for cluster monitoring configuration. The check is not applicable.
Single Node Openshift (Sno) Test Compatibility	✅ Passed	No new Ginkgo e2e tests are added in this PR. The changes only modify API type definitions and CRD schemas for Prometheus retention configuration. The SNO compatibility check is not applicable.
Topology-Aware Scheduling Compatibility	✅ Passed	PR only modifies API type definitions and CRD schemas for Prometheus retention; no deployment manifests, operator code, or scheduling constraints are added or modified.
Ote Binary Stdout Contract	✅ Passed	PR modifies only Kubernetes API type definitions and CRD YAML manifest; no executable binary code, test binaries, or stdout-writing process-level functions are present.
Ipv6 And Disconnected Network Test Compatibility	✅ Passed	No Ginkgo e2e tests are added in this PR. Changes are limited to Go type definitions and CRD YAML files for Prometheus monitoring configuration, making this check inapplicable.
No-Weak-Crypto	✅ Passed	PR only modifies Prometheus retention configuration schema; contains no cryptographic code, crypto imports, or security operations subject to the weak-crypto check.
Container-Privileges	✅ Passed	Check is not applicable to this PR; no container/pod security configurations exist in the monitoring API schema changes reviewed.
No-Sensitive-Data-In-Logs	✅ Passed	The PR modifies only Prometheus retention configuration types and CRD schemas. No logging code is introduced, and no sensitive data (passwords, tokens, keys, PII) is logged or exposed in the changes.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[marioferh]

@simonpasquier @danielmellado

[danielmellado]

While this should be ok not having a tombstone as we're just loosening limits, the openapi.json diff includes unrelated changes (NetworkObservabilitySpec, registries descriptions, softirqs collector, KMS default) which are likely from other PRs that merged into master since the branch was created. Could you rebase on latest master so the openapi.json diff only shows the PR changes? This would avoid confusion during review and prevent accidentally shipping stale generated data.

[danielmellado]

Hey @marioferh, as commented before we go with just removing the max constraints, I think we should also address the granularity issue that Ayoub and Simon raised.

Right now durationInDays: int32 can't express sub-day retention like 15h or 24h (which is the default for UWM), and sizeInGiB: int32 can't express sub-GiB sizes like 500MiB. The CMO ConfigMap and the Prometheus Operator CRD both accept freeform Prometheus duration/size strings, so customers migrating from retention: "15h" would have no equivalent in the new CRD.

I think we should push back on the integer-only convention for these two fields and switch to string fields with the same Prometheus pattern validation that the Prometheus Operator CRD uses:

duration: string with pattern ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$
size: string with pattern ^(0|([0-9]*[.])?[0-9]+((K|M|G|T|E|P)i?)?B)$
This way CMO can pass the values straight through to the Prometheus CR without any lossy conversion, and all existing ConfigMap values remain valid. The patterns are well-defined and fully validatable at admission time, so we're not losing any safety vs. int32.

@everettraven I know we discussed avoiding duration strings back in #2653, but this comes from the Prometheus Operator CR itself, which validates retention with ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$ and size with (^0|([0-9]*[.])?[0-9]+((K|M|G|T|E|P)i?)?B)$ (https://github.com/openshift/cluster-monitoring-operator/blob/0ce8e4c0bce00381802c25d92ed238775139ee82/assets/monitoring-operator/00_0prometheus-custom-resource-definition.yaml). CMO passes these values straight through to the Prometheus CR, so using durationInDays: int32 and sizeInGiB: int32 introduces a lossy conversion layer that breaks existing configs like retention: "15h" or retentionSize: "500MiB" that customers can set today in the ConfigMap. Would you be open to using string fields with these same Prometheus patterns for this specific case? They're fully validatable at admission time via the regex, and since this is v1alpha1 behind a feature gate we can still change it.

Worst case, as a fallback, we may set hours, but wanted to get back your thoughts on this! Thanks!

[everettraven]

Hey @marioferh, as commented before we go with just removing the max constraints, I think we should also address the granularity issue that Ayoub and Simon raised.

Right now durationInDays: int32 can't express sub-day retention like 15h or 24h (which is the default for UWM), and sizeInGiB: int32 can't express sub-GiB sizes like 500MiB. The CMO ConfigMap and the Prometheus Operator CRD both accept freeform Prometheus duration/size strings, so customers migrating from retention: "15h" would have no equivalent in the new CRD.

I think we should push back on the integer-only convention for these two fields and switch to string fields with the same Prometheus pattern validation that the Prometheus Operator CRD uses:

duration: string with pattern ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$ size: string with pattern ^(0|([0-9]*[.])?[0-9]+((K|M|G|T|E|P)i?)?B)$ This way CMO can pass the values straight through to the Prometheus CR without any lossy conversion, and all existing ConfigMap values remain valid. The patterns are well-defined and fully validatable at admission time, so we're not losing any safety vs. int32.

@everettraven I know we discussed avoiding duration strings back in #2653, but this comes from the Prometheus Operator CR itself, which validates retention with ^(0|(([0-9]+)y)?(([0-9]+)w)?(([0-9]+)d)?(([0-9]+)h)?(([0-9]+)m)?(([0-9]+)s)?(([0-9]+)ms)?)$ and size with (^0|([0-9]*[.])?[0-9]+((K|M|G|T|E|P)i?)?B)$ (https://github.com/openshift/cluster-monitoring-operator/blob/0ce8e4c0bce00381802c25d92ed238775139ee82/assets/monitoring-operator/00_0prometheus-custom-resource-definition.yaml). CMO passes these values straight through to the Prometheus CR, so using durationInDays: int32 and sizeInGiB: int32 introduces a lossy conversion layer that breaks existing configs like retention: "15h" or retentionSize: "500MiB" that customers can set today in the ConfigMap. Would you be open to using string fields with these same Prometheus patterns for this specific case? They're fully validatable at admission time via the regex, and since this is v1alpha1 behind a feature gate we can still change it.

Worst case, as a fallback, we may set hours, but wanted to get back your thoughts on this! Thanks!

We strongly encourage not using durations in favor of a unit-based integer, but if you folks feel strongly that this is the best UX for your end-users I'm not going to prevent you from doing so. I'm not the expert on your user base here, you folks are :).

That being said, if the concern is solely that you cannot reasonably express smaller unit values that are likely to occur, you can change the unit you are using for the integer field to the smallest reasonable unit. For example, durationInSeconds and sizeInBytes.

Even if this is in v1alpha1 we still have to abide by tombstoning rules if we change the serialization of the fields unless we want to move this to a v1alpha2.

[coderabbitai]

Actionable comments posted: 4

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@config/v1alpha1/types_cluster_monitoring.go`:
- Around line 2238-2247: The public JSON name was changed and will break
existing manifests; restore wire compatibility by keeping the original JSON
field name for the struct field DurationInHours (e.g. change the tag to
`json:"durationInDays,omitempty"` so existing persisted objects continue to
round-trip) and add a deprecation/kubebuilder comment on the field explaining it
represents hours (or add a new explicitly named field for hours in a new API
version/conversion path if you intend to support both); update the struct field
DurationInHours and its kubebuilder tags accordingly and ensure any API
docs/comments reflect the restored wire name and units.
- Around line 2243-2255: Remove the hard upper-bound metadata and docs for
retention and storage fields: delete the `+kubebuilder:validation:Maximum=87600`
and the "Maximum value is 87600 hours..." doc line for the DurationInHours field
(symbol: DurationInHours) and delete the
`+kubebuilder:validation:Maximum=2147483647` and the "Maximum value is
2147483647 GiB..." doc line for the sizeInGiB field (symbol: sizeInGiB); keep
the Minimum validations and the rest of the comments intact so the API no longer
advertises or enforces an upper bound.

In
`@payload-manifests/crds/0000_10_config-operator_01_clustermonitorings.crd.yaml`:
- Around line 4136-4147: The CRD change renamed the serialized retention field
in v1alpha1 from spec.prometheusConfig.retention.durationInDays to
durationInHours which will break existing stored objects and clients; revert the
rename in the v1alpha1 CRD so the field remains durationInDays, restore its
original description/format/minimum/maximum, and instead introduce any new
representation by adding a new API version (e.g., v1beta1) with durationInHours
plus proper conversion webhooks or conversion funcs and tombstoning logic to
translate between durationInDays and durationInHours during reads/writes.
- Around line 4143-4145: The CRD still enforces a 10-year ceiling via the schema
entry "maximum: 87600"; remove that maximum constraint so the API no longer
rejects larger retention values at admission time. Locate the clustermonitorings
CRD schema where "format: int32" and "maximum: 87600" appear (the
retention/hours field) and delete the "maximum: 87600" line (and any equivalent
numeric maximums on related retention fields) so the property is unconstrained
by the CRD schema. Leave the type/format (e.g., "type: integer" / "format:
int32") intact if needed, and run schema validation tests to ensure no other
admission-side maximums remain.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository YAML (base), Central YAML (inherited)

Review profile: CHILL

Plan: Enterprise

Run ID: 003efd47-3c8c-4d49-aac1-a3b47834cee3

📥 Commits

Reviewing files that changed from the base of the PR and between 03a1de9 and 71e6477.

⛔ Files ignored due to path filters (5)

• config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings.crd.yaml is excluded by !**/zz_generated.crd-manifests/*
• config/v1alpha1/zz_generated.featuregated-crd-manifests/clustermonitorings.config.openshift.io/ClusterMonitoringConfig.yaml is excluded by !**/zz_generated.featuregated-crd-manifests/**
• config/v1alpha1/zz_generated.swagger_doc_generated.go is excluded by !**/zz_generated*
• openapi/generated_openapi/zz_generated.openapi.go is excluded by !openapi/**, !**/zz_generated*
• openapi/openapi.json is excluded by !openapi/**

📒 Files selected for processing (2)

• config/v1alpha1/types_cluster_monitoring.go
• payload-manifests/crds/0000_10_config-operator_01_clustermonitorings.crd.yaml

[danielmellado]

@everettraven besides the override on the CI, would we need to tombstone this due to the filed removal? (change, technically, but anyway). Thanks!

[simonpasquier]

We do not have enough customer or operational data to pick defensible API-level ceilings

We should collect data via telemetry in the first place.

[simonpasquier]

We strongly encourage not using durations in favor of a unit-based integer, but if you folks feel strongly that this is the best UX for your end-users I'm not going to prevent you from doing so. I'm not the expert on your user base here, you folks are :).

What was proposed is to adopt duration format already supported by the upstream Prometheus operator CRDs. These CRDs aren't perfect but they have been there for many years and I bet that our users are familiar with the format.

[coderabbitai]

Actionable comments posted: 0

[marioferh]

after the discussion changed to string to maintain compatibly with Prometheus upstream even is string are not recommended for API

[danielmellado]

We strongly encourage not using durations in favor of a unit-based integer, but if you folks feel strongly that this is the best UX for your end-users I'm not going to prevent you from doing so. I'm not the expert on your user base here, you folks are :).

What was proposed is to adopt duration format already supported by the upstream Prometheus operator CRDs. These CRDs aren't perfect but they have been there for many years and I bet that our users are familiar with the format.

I agree here with Simon and Mario in that the Promotheus CRD should serve as inspiration for this to maintain the base user familiarity with them. @everettraven could we make a compromise and keep that for now? besides your comment there, could you PTAL? Thanks!

[everettraven]

/lgtm

[openshift-merge-bot]

Scheduling tests matching the pipeline_run_if_changed or not excluded by pipeline_skip_if_only_changed parameters:
/test e2e-aws-ovn
/test e2e-aws-ovn-hypershift
/test e2e-aws-ovn-hypershift-conformance
/test e2e-aws-ovn-techpreview
/test e2e-aws-serial-1of2
/test e2e-aws-serial-2of2
/test e2e-aws-serial-techpreview-1of2
/test e2e-aws-serial-techpreview-2of2
/test e2e-azure
/test e2e-gcp
/test e2e-upgrade
/test e2e-upgrade-out-of-change
/test minor-e2e-upgrade-minor

[everettraven]

Fields intentionally removed and appropriately tombstoned. This is a v1alpha1 API so there are no supported upgrade paths for this API, making removal OK here.

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: everettraven

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [everettraven]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[openshift-ci]

@everettraven: Overrode contexts on behalf of everettraven: ci/prow/verify-crd-schema, ci/prow/verify-crdify

Details

In response to this:

Fields intentionally removed and appropriately tombstoned. This is a v1alpha1 API so there are no supported upgrade paths for this API, making removal OK here.

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[danielmellado]

/retest-required

[marioferh]

/test e2e-azure

[marioferh]

/verified by @marioferh

[openshift-ci-robot]

@marioferh: This PR has been marked as verified by @marioferh.

Details

In response to this:

/verified by @marioferh

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

[openshift-merge-bot]

/retest-required

Remaining retests: 0 against base HEAD 1194f4c and 2 for PR HEAD b4b461b in total

[danielmellado]

we may need to tombstone this as well

----- config.openshift.io ----- v1alpha1 ----- config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings.crd.yaml -----
----- SCHEMACHECK -----
ERROR: NoFieldRemoval: crd/clustermonitorings.config.openshift.io version/v1alpha1 field/^.spec.prometheusConfig.retention.durationInDays may not be removed
ERROR: NoFieldRemoval: crd/clustermonitorings.config.openshift.io version/v1alpha1 field/^.spec.prometheusConfig.retention.sizeInGiB may not be removed

[marioferh]

we may need to tombstone this as well

----- config.openshift.io ----- v1alpha1 ----- config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings.crd.yaml ----- ----- SCHEMACHECK ----- ERROR: NoFieldRemoval: crd/clustermonitorings.config.openshift.io version/v1alpha1 field/^.spec.prometheusConfig.retention.durationInDays may not be removed ERROR: NoFieldRemoval: crd/clustermonitorings.config.openshift.io version/v1alpha1 field/^.spec.prometheusConfig.retention.sizeInGiB may not be removed

but this is already override:

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

[openshift-ci]

@marioferh: marioferh unauthorized: /override is restricted to Repo administrators, approvers in top level OWNERS file, and the following github teams:openshift: openshift-release-oversight openshift-staff-engineers openshift-sustaining-engineers.

Details

In response to this:

we may need to tombstone this as well

----- config.openshift.io ----- v1alpha1 ----- config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings.crd.yaml ----- ----- SCHEMACHECK ----- ERROR: NoFieldRemoval: crd/clustermonitorings.config.openshift.io version/v1alpha1 field/^.spec.prometheusConfig.retention.durationInDays may not be removed ERROR: NoFieldRemoval: crd/clustermonitorings.config.openshift.io version/v1alpha1 field/^.spec.prometheusConfig.retention.sizeInGiB may not be removed

but this is already override:

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[everettraven]

Tombstoning has been done correctly. The override got dropped because something else merged in the meantime.

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

[openshift-ci]

@everettraven: Overrode contexts on behalf of everettraven: ci/prow/verify-crd-schema, ci/prow/verify-crdify

Details

In response to this:

Tombstoning has been done correctly. The override got dropped because something else merged in the meantime.

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[openshift-merge-bot]

/retest-required

Remaining retests: 0 against base HEAD d3390bd and 1 for PR HEAD b4b461b in total

[openshift-merge-bot]

/retest-required

Remaining retests: 0 against base HEAD 5d68333 and 0 for PR HEAD b4b461b in total

[openshift-merge-bot]

/hold

Revision b4b461b was retested 3 times: holding

[everettraven]

Re-adding overrides and cancelling the automatic hold.

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

/hold cancel

[openshift-ci]

@everettraven: Overrode contexts on behalf of everettraven: ci/prow/verify-crd-schema, ci/prow/verify-crdify

Details

In response to this:

Re-adding overrides and cancelling the automatic hold.

/override ci/prow/verify-crdify
/override ci/prow/verify-crd-schema

/hold cancel

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[openshift-ci]

@marioferh: all tests passed!

Full PR test history. Your PR dashboard.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.