☁️🚀 Migrating to Confluence Cloud: What Unexpected Change Surprised You Most?

[Sweetlana Portnaya_Seibert]

Hi Atlassian Community 👋

Many teams approach a move from Confluence Data Center to Cloud thinking it’s mostly a technical migration — move pages, migrate spaces, done. But in reality, Cloud migration often comes with hidden changes that only surface after go-live.

Based on a recent deep dive published on the Seibert Products blog about Confluence Cloud Migration: What Changes (Pricing, Features & Apps), one thing became very clear:
👉 the biggest surprises are rarely about moving content, but about how people experience Confluence afterward.

Some common “wait… what?” moments teams run into:

• Custom CSS, UI tweaks, or user macros no longer behaving the same (or not existing at all)

• Marketplace apps working differently — or requiring manual reconfiguration

• A faster release cadence introducing new features (and UI changes) more frequently than expected

• Shifts in governance, permissions, and how navigation or hubs are structured

• Discovering that Cloud offers powerful new features (automation, analytics, AI search), but requires rethinking old workflows

So I’m curious 👇

What was the most surprising “hidden” change you encountered after moving to Confluence Cloud — good or bad?
Was it apps, customization, user adoption, pricing, governance, or something else entirely?

Looking forward to learning from your experiences! 😊

[Anahit Sukiasyan]

Thanks for sharing it here as well, @Sweetlana Portnaya_Seibert! Really enjoyed putting it together with you and looking forward to seeing the community’s thoughts and experiences. 😊

[Calogero Bonasia]

Atlassian has been promoting Jira Cloud migration for years as a natural evolution for organizations using Data Center, promising continuous innovation, reduced infrastructure costs, and simplified management.

However, the operational reality emerging from concrete experiences reveals a significantly more problematic picture, characterized by critical issues that compromise users' daily efficiency.
The first critical aspect concerns system performance. Numerous users document systematic slowdowns with loading times regularly reaching 13-30 seconds for ordinary operations, 100% CPU consumption during normal use, and frequent timeouts.

These problems have persisted from 2019 to 2025, independent of client configuration, and affect organizations of varying sizes, suggesting structural limitations of the Cloud architecture rather than isolated anomalies.

Particularly problematic is the loading of issues (4-30 seconds), board updates, and simple navigation between screens—operations that in Data Center environments required fractions of a second.
Custom field management represents a second significant criticality. The Migration Assistant tends to duplicate custom fields even when they possess identical properties, generating multiple entities marked with progressive suffixes ("migrated", "migrated 2", "migrated 3").

This proliferation causes malfunctions in automation rules, user confusion, and requires complex post-migration cleanup interventions through dedicated scripts. Documented cases report fields that lose their context during transfer, becoming inaccessible via API and invalidating existing JQL queries.

The identical naming of previously distinct fields (such as "Start date" and "Start Date") generates additional operational problems, making it impossible to distinguish them in configurations.
Functional limitations complete the critical picture. Jira Cloud imposes an insurmountable limit of 32,767 characters for descriptions and comments, non-modifiable as in Data Center, causing errors during migration when content exceeds this threshold.

Since December 2024, Atlassian has also introduced a limit of 10,000 worklogs per issue, forcing organizations using Tempo Timesheets to implement aggregation solutions or continuously create new issues for recurring activities, complicating reporting and traceability.
BigPicture users report specific synchronization issues in Cloud environments: updates that require minutes or even days to propagate, changes that disappear after refresh, the need to manually rebuild structures to make changes persistent.

The bidirectional synchronization between BigPicture and Jira proves particularly unstable, requiring frequent manual re-sync interventions or cache clearing that temporarily blocks the application.

Custom workflows that don't migrate automatically, Marketplace applications requiring manual reconfiguration despite the existence of Cloud versions, and insufficient database connection pools during migration complete a concerning panorama.
After several years of commercial push toward Cloud, structurally disconnected elements persist that compromise daily operational experience, raising legitimate questions about the platform's actual maturity compared to the expectations generated by the vendor.