Created attachment 1556353 [details] RobinStorage Description Description of problem: Use correct and distinctive categories for different operators, also we need update Description font for Robin Storage Version-Release number of selected component (if applicable): 4.0.0-0.nightly-2019-04-18-190537 console commit: io.openshift.build.commit.id=8a2df3e00d50c88f630e329159e6bcd98f6b2767 io.openshift.build.commit.url=https://github.com/openshift/console/commit/8a2df3e00d50c88f630e329159e6bcd98f6b2767 How reproducible: Always Steps to Reproduce: 1. cluster admin logs into admin console and visit Catalog -> Operator Hub page $ oc get packagemanifests robin-operator -n openshift-marketplace -o yaml | grep categories categories: Database,BigData,Storage $ oc get packagemanifests spark-gcp -n openshift-marketplace -o yaml | grep categories categories: Big Data Actual results: 1. There are two similar Categories: Big Data and BigData, user can't view the differences between these categories Expected results: 1. Use distinctive categories for different operator Additional info: This may need operator author to change the configurations and not console issue
Console is using the metadata from the package manifest. I think the best solution is to align the categories there.
The PackageManifest is just passing through the annotations defined by OperatorHub. I couldn't track down the docs for the current set of Categories, so I'm moving this to the OperatorHub component
Accepted categories are listed here https://github.com/operator-framework/community-operators/blob/master/docs/required-fields.md#categories. I will create a PR to map the categories from packagemanifest to this set of categories. If none of the categories listed in the packagemanifest are on the above list, the operator will be categorized as "other" in the UI. Does that sound good?
Actually, this is an error with the CSV so I will submit PRs against the community-operators repo to fix these CSVs
I think it would be worthwhile having an agreed list of categories in a file that multiple components can draw from to make validation. Long-term we should have a shared library that multiple components can use to determine the validity of an Operator and decide whether to display it or not.
operator-courier should know what the valid categories are and emit a linting error if the category is not valid. Apparently it already does this with the validate-ui flag, but were asked to not use that flag since it was tuned for operatorhub.io. We can enable using that flag for testing operators again, we can pull that test into the core set of linting tests, or we can add an additional flag specifically for embedded operatorhub.
The category of operator "AppDynamics ClusterAgent" is `Application Monitoring`, suggest it to be `Monitoring`; the category of "Tufin Orca Operator" if `Security Policy Management`, suggest it to be `Security`. And all the categories of our default opsrc should belong to the list of the recommand categories in https://github.com/operator-framework/community-operators/blob/master/docs/required-fields.md#categories.
This should not be attached to a specific openshift release. There is a process in community-operators for setting these categories.