Nexus Repository Community Edition allows 40,000 components or 100,000 requests a day. Above that, nobody can publish anything new until you are back under — or you buy Pro. We find out what fills your count and pick the cheapest way back: clean up, split the load, or move out.
Over 40,000 components or 100,000 requests a day, Nexus CE blocks new uploads.1
Measure what fills the count, then clean up, split or move — whichever is cheapest for you.
Builds can publish again, without a Pro licence, and you stay under the limit.
From Sonatype's own documentation:1
We recommend one after measuring your instance. Sometimes the first is all you need.
| Way | What we do | Fits when |
|---|---|---|
| 1. Clean up and stay | Cleanup policies for snapshots and old releases, unused repositories removed, request spikes traced to the CI job causing them | Your own artefacts are mostly old versions nobody needs |
| 2. Split the load | Container images and Helm charts moved to Harbor, libraries you publish moved to the GitLab Package Registry; Nexus CE keeps what it does well, for example proxying Maven or npm | Containers or a few big projects push you over the limit |
| 3. Move out | Full migration to Harbor and GitLab registries, Nexus switched off — see Nexus / Artifactory migration | You are far over the limit, or want one platform less |
What makes up your component count and request volume: by repository, by format, by age.
Written and applied with you, starting with conservative rules, after a full backup.
The CI jobs or scripts that hammer Nexus, and how to fix them.
Usage dashboard green again, or a clear plan for the split if cleaning is not enough.
An alert before you reach the limit again, so it never blocks a release by surprise.
Shown for a typical mid-size setup. Larger platforms take longer; the scope and price are fixed on the first call.
Full backup first. Then the usage breakdown and the request logs.
Cleanup policies applied repository by repository; we watch the counts fall.
Alerts set up. If cleaning is not enough, a fixed quote for the split.
Fixed price for measuring, cleaning up and guard rails. The split, if needed, gets its own fixed quote.
Only after a full backup, with rules you approve, and starting with the oldest snapshots. Nothing your customers or builds still use is touched.
We measure exactly what makes up your count on day one, so you see which repositories fill it before anything is changed.
If Pro features such as high availability matter to you, it can be the right answer. Many teams only need to publish again — that is cheaper to fix.
Not yet in production: GitLab's Maven virtual registry is beta and others are not released. We keep proxies on Nexus CE or Harbor for now.
Components and requests per day — enough to tell you which of the three ways fits.