ACCESS System Status News Policy, Tools, and Best Practices
This document outlines status news policy, tools, and best practices for ACCESS integrated infrastructure. As of October 2024, ACCESS Operations applies these elements to infrastructure it operates. ACCESS Operations plans to share this document with the ACCESS Executive Council for review, revision, and potential adoption by all ACCESS projects, and with the Resource Provider Forum for potential adoption by all ACCESS resource providers.
Motivation
ACCESS researchers, resource providers, projects, and affiliated service providers (such as science gateways), rely on ACCESS infrastructure provided by resource providers, projects, and partners or vendors (Google, Globus, Atlassian, etc.). Timely news about planned or unplanned outages and significant configuration changes improves researcher experience, streamlines ACCESS support processes, and enables ACCESS infrastructure providers to support their customers when other infrastructure that they rely on has status events. All ACCESS ecosystem participants, whether they use or provide infrastructure, benefit from timely status news.
Status News Policy
ACCESS integrated infrastructure providers should publish all significant status news as early as possible through ACCESS channels. ACCESS community members interested in status news should be able to visit public websites to view current, upcoming, and recent status news, and should be able to receive timely status news emails.
Definitions:
Infrastructure scope: This policy applies to infrastructure integrated by ACCESS resource providers, ACCESS projects, and partners or vendors using ACCESS infrastructure integration roadmaps. Infrastructure visible to ACCESS researchers, between ACCESS projects, and to ACCESS developers, is in scope. Infrastructure from 3rd parties and commercial vendors with an active contract to provide service to ACCESS is in scope. Infrastructure not directly visible/usable outside a single resource provider or project is not in scope. The terms “system” and “infrastructure” are synonymous in this document.
Types of news: Status news types include full outages, partial outages, degraded performance, new system announcement, retirement, and significant configuration changes. Partial outages affect a subset of users or a subset of functionality.
Timing for news: Planned status events should be published at least 2 weeks ahead of time. Unplanned status events should be published as soon as possible. Retirements should be published at least 6-months ahead of time, or a year when researchers may need more time to migrate to other infrastructure.
Significant: Status changes are significant if they last longer than 15 minutes or are likely to affect more than a dozen people. Deciding what is significant enough to publish is a judgment call by the infrastructure provider. The pre-production or testbed status of infrastructure is a factor for the provider to consider. All retirements are significant.
Online news: Status news must be public and viewable through one or more portals. Members of the ACCESS community should be able to easily view current, planned, and recent status news. Online news is important for individuals that aren’t getting emails, for staff providing support for infrastructure that they don’t directly use, and for anyone investigating recent application, software, or service failures.
Opt-in emails: ACCESS community members should be able to opt-in to proactively receive status news emails about ACCESS infrastructure of their choosing.
Publisher: Status news should be published by the ACCESS provider or integrator. For infrastructure from 3rd parties and commercial vendors, the ACCESS organization that integrated it and supports it in ACCESS should cross-publish significant status news through ACCESS and should reference vendor published status news.
Validation: Authorized individuals (CyberSecurity staff and leadership) must be able to verify the authenticity of published news by reviewing logs showing who published the news.
Authentication and Authorization: News can only be published by ACCESS authenticated and authorized individuals.
Status News Tools and Services
Publishing News
ACCESS integrated infrastructure providers MUST publish or cross-post significant status news to the ACCESS System Status News service by either:
- Entering it by hand in the Operations Portal
- Publishing it by API to the ACCESS Operations Information Sharing Platform
Instructions on how to publish and update status news by hand is in the Operational Status Communication integration roadmap task.
Viewing News
ACCESS Operations and Support provide system status news views in their portals:
ACCESS Support provides affinity groups where members of the ACCESS community can opt-in to receive emails for affinity group related news. Affinity group owners can request that status news about infrastructure of interest to their group be visible in their affinity group, and that this news be emailed to subscribed members of their affinity group:
ACCESS Operations provides APIs for accessing infrastructure status news entered in the Operations portal and published by API:
Status News Best Practices
- Links in published status news and emails must point directly to the destination site and not use click tracking services.
- Infrastructure providers may have local status communications methods independently of ACCESS.
Feedback
Please submit questions or suggestions about this document by opening an operations request and selecting the "Operations: Other" issue type.