How to Bulk-Deploy and Schedule WordPress Content Without Timing Out
You select 40 posts on the Dashboard and click Deploy once, expecting one push per post to run back to back inside that single page load. For a handful of posts that works fine. For 40, the request itself becomes the problem — PHP's execution time limit, the web server's own timeout, or just the browser tab sitting there waiting, long before the last few posts even get a turn. "Select many, deploy at once" and "run each one inside the request that asked for it" turn out to be two different things, and only one of them scales.
A bulk deploy and a scheduled one are the same thing underneath
Selecting 40 posts and clicking Deploy now, and scheduling one post to push next Tuesday at 9am, look like different features from the toolbar. They aren't. Both just insert a row into a queue table — post_id set for a push, source_uuid set for a pull from a connection — with a timestamp for when it's due. "Deploy now" is a row whose timestamp is already due. "Schedule once" is a row whose timestamp is in the future. "Repeat weekly" is the same row with a recurrence attached. There's no separate code path for bulk versus scheduled versus recurring — one enqueue call, different values.
A background worker drains that table in small batches rather than all at once, so a bulk selection of 40 doesn't become one long synchronous request either — it becomes several short ones, a handful of posts at a time, seconds apart. Each item in the batch goes through the exact same push or pull logic a single manual deploy uses — same validation, same reference resolution, same conflict detection. Nothing about going through the queue is a lighter-weight or less-checked version of a manual deploy.
What a recurring schedule actually keeps
A "repeat weekly" item doesn't live as one row that gets reset and reused every week. Each run that completes successfully creates a fresh row for the next occurrence, rather than rewinding the one that just finished. That means the history for a recurring push isn't a single "this schedule exists" record — it's a full list of every individual run, each with its own status, its own attempt count, its own error if it had one. If run #6 of a weekly push failed and run #7 succeeded, both show up as separate, inspectable entries, not one row that silently overwrote itself.
When the destination isn't reachable
A connection being down when a scheduled push fires doesn't fail silently and it doesn't get skipped — it retries with increasing delay, three times, before giving up for good. A handful of error types never get retried at all (a bad package, a failed authentication, a post type the destination doesn't support), because retrying something that can't succeed just burns through the backoff schedule for nothing. An unreachable connection isn't one of those — it's treated as the kind of failure that might resolve itself, so it gets retried before anything gives up.
One detail worth knowing if you're relying on a recurring schedule: the next occurrence only gets created after a run succeeds. If a recurring push exhausts its retries and ends in permanent failure, that chain stops — it doesn't quietly skip ahead and try again at the next interval. A failed recurring item sits there as a visible failure on the Queue page, waiting for a person to look at it, rather than disappearing into "well, it'll try again next week anyway."
The part that's easy to get wrong
Every connection has a stored default direction — push or pull — set when it was created. The obvious way to decide what a queued item should do is check that stored direction. That was the actual first implementation, and it was wrong: nothing stops the Dashboard from pushing through a connection whose stored default happens to be "pull," since the direction a connection was created with was never meant to be exclusive. A scheduled push enqueued against one of those connections got routed down the pull branch anyway, which looked for a source object to pull, found none, and failed instantly with an error about having nothing to pull — on a push the admin had very deliberately scheduled.
The fix was to stop asking the connection what direction it defaults to, and instead look at which field the queued item itself actually populated — a push item has a post ID, a pull item has a source UUID, and that's unambiguous regardless of what the connection's stored default happens to be. The same exact mistake turned up in the REST automation endpoint a day later, fixed the same way — a reminder that a shortcut like "just check the connection's direction" tends to be wrong everywhere it's used, not just once.
What PostDeploy does about it
Bulk selection lives on the same toolbar whether you're pushing your own posts from the Dashboard or pulling objects in from a connection on the Remote Content screen — pick posts (or remote objects), optionally set a date and a repeat interval, and the rest is the queue described above. The Queue page shows every pending, completed, and failed item with its connection, its next attempt time, and its error if it has one, and lets you cancel or reschedule anything still pending.
Because WP-Cron only ever fires off the back of an actual page load on the site — not on its own, on a fixed clock — a low-traffic site can have a due item sit for a while with nothing actually wrong. That's a general WP-Cron characteristic, not specific to bulk/scheduled deployment; the auto-push on publish post covers it in more depth. The Queue page's "Run Queue Now" button exists specifically to sidestep it — it runs the worker immediately, synchronously, rather than waiting for WordPress's own scheduler to get around to it. And a pull that did fail and run out of automatic retries isn't stuck forever either — WP-CLI's retry command can re-run that specific deployment directly, outside the queue entirely. A failed push doesn't have that same command; push it again from the source instead.
Try it
Bulk and scheduled deployment is a Pro feature. PostDeploy Free moves one Post, Page, or custom post type at a time, on demand. PostDeploy Pro adds selecting many at once, scheduling them for later, and repeating them on a recurring interval — for both pushing and pulling.