Drupal 10 and the Luxury of a Boring Upgrade

Drupal 10 shipped on December 14, and the January conversations it has generated are the kind we spent a decade waiting to have. Clients asking, without anxiety, when should we schedule the update? Not ‘how do we survive the migration.’ Not ‘what will this cost us.’ Scheduling questions, of the kind one asks about routine maintenance, because that is what major-version upgrades on this platform now are.

The release itself is substantial where users touch it. CKEditor 5 replaces the decade-old editor with a genuinely modern writing surface. The Olivero and Claro default themes bring front and admin experiences that no longer need immediate replacement. Symfony 6 refreshes the foundation underneath. And it is familiar where operators touch it. Built from Drupal 9’s final minor with deprecated code removed, exactly per the model we celebrated in 2020.

That model has now delivered twice, which converts it from promising precedent to demonstrated practice. So this post is short on drama and long on the two things the moment actually calls for. The scheduling mechanics for the 9 sites. And the much sharper talk this release forces on the audience that still has a real decision. 10’s arrival and Drupal 7’s November end-of-life landed in the same season, and only one of those dates can move. It is not the November one.

The mechanics, briefly, for the 9 sites

The upgrade path, as our maintenance book is running it this quarter. The deprecation scan, the same tooling and habit the 9 transition installed, against your custom code. The module-readiness check: the contributed ecosystem crossed ahead of release day again, with the long tail closing weekly. The 9.5 stepping stone: sites should be current on 9’s final minor before jumping, which the maintenance calendar should have handled already. Then the update itself through the standard staging path, annotated as always.

Sizing, honestly: an afternoon for the well-kept sites, a sprint for the larger and moderately customized ones. The effort sits almost entirely in whatever old-API residue the scans find, and in CKEditor 5’s arrival. That is the one user-facing change worth planning around. Custom editor configurations and plugins from the CKEditor 4 era need re-mapping. Editors deserve a heads-up and a short walkthrough. The new editor is better in ways they will notice at once, and different in ways they will notice first.

That is the whole list. The saved anxiety is real, purchased by the release-engineering practice we documented in 2020. Our suggestion for spending it is the next section, because the sites with an actual decision this year are not the ones running 9.

The audience with a real decision: Drupal 7, final year

• The convergence is the message: Drupal 10 arrives as Drupal 7 enters its final supported year. November 2023, per last April’s extension, with the community signaling firmly that the extensions are done. The migration destination is now maximally modern and the deferral runway is minimally short. Together, those facts make 2023 the unavoidable action year. Sites that heard last April’s advice as ‘someday’ should hear this one as ‘now.’

• The migration math has less runway: the April subtraction (deadline minus build, buying, and decision time) ran against eighteen months. It now runs against ten. For institutional speeds, that means the decision phase must be this quarter and the procurement by summer. Even then, the honest schedule for many sites crosses the deadline mid-project, which is precisely what the extended-support bridge exists for. Purchased with a project plan attached, not instead of one.

• The triangle stands, re-weighted: rebuild to modern Drupal, replatform elsewhere, or bridge-and-decide. The corners are unchanged from April, and so are the criteria. The only re-weighting is time. Options that were choices in April are becoming consequences by autumn, and we run the calendar-candor category precisely so its readers stay on the choice side of that line.

The evaluator’s note, which would have sounded implausible in 2011

One paragraph for the platform-selection crowd, because it marks a real change worth recording. Modern Drupal’s editing experience no longer needs an apology. Readers with long memories, or 7-era scars, will remember when Drupal evaluations meant conceding the authoring experience and arguing the trade. The structural power, the taxonomy, the language machinery. Bought at the price of editor suffering we spent years easing with admin themes and training.

Together, CKEditor 5, the Claro admin experience, and the layout tooling have closed that gap. Our write-ups this year call the authoring tier competitive rather than tolerable. The usability objections of the 7 era are, on current evidence, answered.

That matters for the 7-migration evaluations above. The ‘rebuild to modern Drupal’ corner is a materially better destination than the one many holdouts last evaluated. And it matters for the general selection frameworks. The platform’s old weakness has been engineered away with the same patience as its upgrade trauma. Evaluations still citing 2015-era authoring complaints are citing a platform that no longer ships. Credit where the decade’s work has earned it.

What repeatability changes

The continuity model has now survived its second full cycle. 8-to-9 proved it possible. 9-to-10 proves it repeatable. Repeatability is the property that changes planning. Institutions can now write Drupal horizons in maintenance terms with two data points of evidence rather than one promise, which is how a release model becomes a total-cost fact.

The pattern this joins is the boring-majors trend across the serious platforms. WordPress’s kept promise on compatibility. The evergreen browsers. The cloud services’ steady delivery. It is our favorite growing-up story. It is the one that pays readers directly. Every platform that learns to change without trauma converts its users’ upgrade budgets into improvement budgets, permanently.

And the luxury named in the title deserves its closing sentence, because it is the actual thesis. Infrastructure should feel boring. Boring is expensive to engineer and cheap to enjoy. A platform that spent a decade earning this January’s mundane scheduling questions has delivered the only kind of excitement we ever wanted from its foundations: none. Schedule the update. Spend the drama budget on the 7 sites, where the calendar is still capable of supplying some, for exactly one more year.