Why scheduled posts fail, and what to do about each cause

Expired connections, rejected media, rate limits, silent tools. The handful of things that actually go wrong, and how to tell them apart.

A scheduled post that does not go out is a specific kind of annoying, because you usually find out late and often by accident. The causes are fewer than you would think, and they are distinguishable if you know what to look for.

The connection expired

By far the most common cause. Every network expires the access it grants a tool. Change your password, revoke an app, lose admin rights on a Page, or simply wait long enough, and the connection stops working.

How to tell: everything for one channel fails while others are fine. The fix: reconnect. The thing to check afterwards: whether the posts queued behind it were held or burned through one at a time at their scheduled slots. A tool that parks them lets you carry on; one that fails them individually leaves you with a column of failures and an empty queue.

The media was rejected

Every network has its own rules about formats, dimensions, duration and file size, and they are not the same. A video that is fine on one is refused by another.

How to tell: posts with media fail, text-only posts succeed. The fix: usually re-encoding to H.264 MP4, which is the safest thing everywhere. Some tools convert automatically, which is worth knowing before you upload forty clips.

You hit a limit you did not know about

Networks cap what an app can do in a window, and some caps are surprisingly low. Instagram counts each story frame against a daily quota. Bluesky limits video uploads per day. These are rarely documented where you would look.

How to tell: it worked earlier and stopped, without anything changing. The fix: wait, and spread posting out rather than batching everything into one hour.

The post was too long for that network

Especially when posting the same text to several networks with different limits — and different counting rules. A post that fits Mastodon's 500 does not fit Bluesky's 300, and a tool that counts to one number is wrong on the other. The per-network breakdown is here.

Something changed on the network's side

Permissions get deprecated, endpoints change, a required field appears. Rare, affects everyone at once, and outside your control. How to tell: it is failing for other people too — check the tool's status page or their social account before debugging your own setup. Meta is the most frequent source of this, and worth reading about separately.

The failure mode that matters most

"The problem is rarely that a post failed. It is that you found out six days later."

Everything above is recoverable in minutes if you know about it. The thing worth checking when evaluating any scheduler is what happens on failure: does it retry transient errors, does it show you the network's actual error message rather than a generic one, and does it tell you unprompted?

A five-minute diagnostic

When something fails and you do not know why, this order narrows it fast:

• Is it one channel or all of them? One channel means the connection. All of them means the tool or the network.
• Is it posts with media, or all posts? Media-only means format or size.
• Did it work earlier today? Yes means you probably hit a limit. No means something changed.
• Does the error name the network? A message from Facebook or Bluesky is actionable. A generic "failed to publish" means your tool is not showing you what it was told, which is its own problem.
• Can you publish the same post manually? If yes, the network is fine and the problem is between you and the tool.

Preventing most of it

• Reconnect channels before they break, roughly every few months, rather than waiting for the failure.
• Keep media in H.264 MP4 and standard image formats. Exotic codecs are the second most common cause and the easiest to eliminate.
• Spread posting out. Batching forty posts into one hour is how you find rate limits.
• Check the queue after any password change on a connected account. That single action invalidates connections on several networks at once.

The test worth doing once

Test it deliberately. Revoke a connection on purpose and see what your tool does — whether it tells you, what it does with the posts queued behind it, and whether reconnecting restores them or you have to rebuild the queue by hand.

It is five minutes and it tells you more than any review will. Everything else on this page is recoverable; a tool that fails silently is the one that costs you a month before you notice.