ElasticPress Hosting

ElasticPress hosting for search that left the database.

Once search runs against an external index, your results are a copy of your content. Copies drift, and a drifted search index is wrong without ever looking broken.

Reindexing completesDrift caughtFallback when unreachable
Batch reindexingDrift checksQuery fallbackDaily backupsFree SSL24/7 support
Built for external search

Two copies of the truth.

Moving search out of MySQL fixes the performance problem and introduces a synchronisation problem in its place.

Reindexing that reaches the last document

A full reindex pushes every post, term and meta field to the index. It runs as a resumable batch, so a large site finishes rather than stopping wherever the timeout landed.

Resumable

Drift between index and database, caught

A failed sync leaves a post that exists but cannot be found. Document counts are compared against the database so a gap is noticed rather than reported by a customer.

Counts compared

A fallback when the index is unreachable

If search depends on a service that is briefly down, the failure should be a slower query rather than an empty results page. Queries fall back to the database instead of returning nothing.

Mapping changes handled as migrations

Altering how a field is indexed usually requires a rebuild, not an edit. Those runs are treated as scheduled work rather than something applied live and hoped over.

Bulk edits that do not flood the index

Changing a term across thousands of posts queues thousands of index updates. Those are batched instead of fired one per save while an editor waits.

Synchronisation

The index is not the source.

Everything that goes wrong here is a divergence: content that was saved and never indexed, or indexed under a mapping that has since changed.

  • Full reindex runs as a resumable batch
  • Index and database counts reconciled
  • Queries fall back rather than returning nothing
  • Bulk edits queued instead of fired per save
Documents
Indexed
184,000
Drift
vs database
0
Fallback
If unreachable
database
Reindex
After mapping change
running
Query time

Fast, and actually complete.

The point of an external index is that a faceted search stops being a meta query. The risk is that it returns quickly and leaves things out.

Resumable
reindex batches
Reconciled
index against database
Graceful
fallback on outage
MagicWP60 ms
Typical shared host1180 ms
MySQL meta query search1450 ms

Illustrative comparison of a faceted search query against an external index. Your numbers depend on document count and facet complexity.

Configuration

What we change for ElasticPress

Search now lives outside WordPress, so every row is about keeping the outside copy in step with the inside one.

SettingWhat we doWhy
Reindex runsResumable batches with progress retainedA full rebuild pushes every document to the index, which on a large site cannot finish inside a request and silently stops part way.
Drift detectionIndex document counts reconciled against the databaseA post that saved but failed to index exists and cannot be found, and nothing about the site looks wrong until somebody searches for it.
Query fallbackFalls back to the database when the index is unavailableAn unreachable search service should degrade to a slower query rather than an empty results page that looks like missing content.
Mapping changesTreated as scheduled rebuilds, not live editsChanging how a field is analysed does not apply to documents already indexed, so results stay inconsistent until everything is rebuilt.
Bulk edit indexingQueued rather than triggered per saveReassigning a term across thousands of posts fires an index update for each one, which blocks the editor and floods the service at the same time.
Plans

Simple, transparent pricing.

Every plan includes free migration, daily backups, SSL and 24/7 support.

MonthlyYearly 2 months free
Starter
For personal sites, blogs, and portfolios.
$20/mo
  • 1 WordPress site
  • 10 GB NVMe disk
  • Free SSL
  • Daily backups
  • One-click deployment
  • Support tickets
Start free trial
Pro★ Most popular
For growing businesses and busy stores.
$80/mo
  • 5 WordPress sites
  • 50 GB NVMe disk
  • Free SSL
  • Daily backups
  • One-click deployment
  • Priority support tickets
Start free trial
Enterprise
For agencies and high-traffic platforms.
$250/mo
  • 20 WordPress sites
  • 200 GB NVMe disk
  • Free SSL
  • Daily backups
  • One-click deployment
  • Dedicated support
Start free trial
FAQ

Questions, answered.

Why can't I find a post that definitely exists?
Almost always index drift — it saved to the database and failed to reach the index. That is why document counts are reconciled here, so a gap gets noticed rather than reported by a visitor.
What happens if the search service goes down?
Queries fall back to the database. It is slower, but an empty results page reads to a visitor as missing content rather than as a temporary outage.
Do I need to reindex after changing a field mapping?
Yes. Documents already indexed keep the old analysis, so results stay inconsistent until a rebuild completes. Those runs are scheduled as batches rather than applied live.
Is ElasticPress worth it on a medium-sized site?
It depends on the queries. Faceted filtering across many meta fields is where a database search collapses; a straightforward keyword search on a few thousand posts usually does not need it.

Keep the index and the content agreeing.

Move the site across — migration is free, reindexing handled.