Epic Goal

• Provide a mechanism to enable upgrades from failed operator installations without manual intervention; i.e. "Fail Forward"

Why is this important?

• Today, when an operator's installation fails due to a defect – e.g. it's bundle has a bad CRD spec – users must manually reconcile k8s resources – e.g. delete the operator – before it can be upgraded to a working version
• When a defect is known and a working version is published, manual intervention becomes unnecessary overhead for both users and authors

Scenarios

1. An issue with the operator's manifests – e.g. a typo – prevents the installation from succeeding
2. The operator's deployment crashes; e.g. bad code

Acceptance Criteria

• Authors can opt-in to allow automatic upgrades of a previous failed installation (pegoncal: removing this from A/C as it got descoped)
• Users can opt-in to automatic upgrades of failed installations for specific operators (if allowed by author)
• Failed installations do not automatically upgrade without user opt-in
• Once automatic upgrades of a failed operator are enabled, no further intervention by the user is necessary to ensure available upgrades are applied

Related Issues

• Addon Operator Incident Report
• OLM Gaps for OpenShift Dedicated
• BZ 2006773
• Managed Services and OLM

Stakeholders

• Service Delivery: nschiede@redhat.com 
• Operator Framework: DanielMesser
• SREP: jaharrin, dofinn, cblecker.openshift
• MTSRE: asegundo+sd-mt-sre

Done Checklist

• CI - CI is running, tests are automated and merged.
• Release Enablement <link to Feature Enablement Presentation>
• DEV - Upstream code and tests merged: <link to meaningful PR or GitHub Issue>
• DEV - Upstream documentation merged: <link to meaningful PR or GitHub Issue>
• DEV - Downstream build attached to advisory: <link to errata>
• QE - Test plans in Polarion: <link or reference to Polarion>
• QE - Automated tests merged: <link or reference to automated tests>
• DOC - Downstream documentation merged: <link to meaningful PR>