"Publishing is slow" is one of the most common complaints I hear from Sitecore content teams. Sometimes a publish takes hours. Sometimes the dialog sits at the same item count forever. The good news is that the causes are predictable.
How Sitecore publishing works, briefly
Publishing copies item versions from the master database to one or more publishing targets, usually web. There are three modes:
- Incremental publish publishes only the items in the publishing queue, meaning items changed since the last publish.
- Smart publish compares revisions between master and the target and publishes only what differs. It checks every item in scope, so it's slower on big trees.
- Republish overwrites everything in scope, whether it changed or not. It's the slowest by far.
After every publish, Sitecore clears caches and updates indexes on each server. That work counts toward "publishing is slow" even though it happens after the copy.
1. Stop the republish habit
In most organizations someone discovered that "republish the whole site" fixes a stale page, and it became routine. A full-site republish on a large tree can take hours and blocks other publishes. Train editors to publish the item they changed, with subitems only when needed, and keep site-level republishes for admins.
2. Clean the queue and event tables
Check row counts in the EventQueue, PublishQueue and History tables of your databases. Millions of rows mean the cleanup agents aren't keeping up. Review the CleanupEventQueue, CleanupPublishQueue and CleanupHistory agent settings in your configuration, and shorten their retention. A one-off cleanup during a quiet window often makes a dramatic difference.
3. Try parallel publishing
Sitecore can publish languages and targets in parallel. Look at the Publishing.MaxDegreeOfParallelism setting and the parallel publishing example config shipped with your version. Test it on a non-production environment first, because it raises the load on SQL.
4. Consider Sitecore Publishing Service
For sites with large content volumes, Sitecore Publishing Service (SPS) replaces the built-in publisher with a separate service designed for bulk publishing. It's much faster for big publishes and doesn't tie up the content management server. It's an extra component to install and patch, so check the compatibility matrix for your Sitecore version and weigh the operational cost.
5. Keep indexing off the critical path
If the web index rebuilds after every publish, publishing appears to hang. The web index should normally use the asynchronous onPublishEndAsync strategy, which updates only changed items. Full rebuilds belong in a scheduled job, not in the publishing flow.
6. Check workflow states
Items that aren't in a final workflow state are skipped silently. The publish "succeeds" but the page doesn't change, which leads to more republishes. Use the Workbox to find items stuck in draft or awaiting approval, and make sure editors know about it.
7. Watch cache clearing on content delivery
After a publish, every content delivery server clears its HTML cache and reloads items. On big sites, many publishes in a short time can keep the caches permanently cold. Batch publishes where you can, for example with a scheduled publish every 15 or 30 minutes instead of dozens of ad-hoc ones.
8. When a publish is stuck
When the dialog stops moving, check these in order:
- The publishing log and the main Sitecore log on the content management server for exceptions.
- Whether another long publish is already running. Publishes queue behind each other.
- The
EventQueuetable size, and whether content delivery servers are processing remote events. - SQL blocking on the master or web database during the publish.
- App pool recycles during the publish, which abort it.
What to do next
Start with the cheapest fixes: retrain editors away from republishing, then clean the tables. Most teams see publishing drop from hours to minutes with those two changes alone. If you're still stuck, I can review your publishing setup as part of a support engagement.
