Understanding Your Usage
As software development is mission-critical to business operations, ensure your Nexus Repository is deployed appropriately for your scale. The usage metrics alert administrators to when it’s time to review their deployment model.
See the Usage Metrics documentation to learn how these are calculated.
The Understand your Usage Banner is removed in Nexus Repository 3.77.0 in favour of a Status Check to inform when Pro users are over the recommended usage for their deployed database.
See the Soft Limits Metrics in Status Checks for details.
Why Did Sonatype Add Usage Warnings for Embedded Databases?
While embedded databases are appropriate for disposable edge caches, test deployments, and low-concurrency team usage, an external PostgreSQL is recommended for more demanding use. This is particularly true for higher-concurrency use, or in any situation where Nexus Repository is the system of record for the components that it stores.
Concurrent loads lead to data integrity problems in OrientDB.
Embedded databases require the system to enter read-only mode to get a consistent backup. Many deployments are not properly backed up, leading to extremely long recovery times in the case of an outage that requires restoring data.
To determine an appropriate deployment model, we’ve created reference architectures providing the deployment recommendations for ranges of anticipated load.
Review the sizing instructions to evaluate your projected load.
Seeing Usage Alerts
If you are seeing usage alerts in your deployment, your scale is exceeding what is appropriate for your deployment. Contact the Sonatype Sales team to upgrade to a Nexus Repository Pro deployment.
When on Nexus Repository Pro
Nexus Repository Pro deployments using an embedded database will see usage alerts when your scale exceed what is appropriate for your deployment.
See Resilient Deployments for recommendations.
Triggers for In-Product Usage Level Warnings
For deployments of Nexus Repository that use an embedded database (OrientDB or H2), in-product warnings appear when usage levels approach or exceed one of the following thresholds:
20,000 requests per day
100,000 components
What Migrating to Pro or PostgreSQL is Not Possible
When changing your deployment model is not possible, we recommend evaluating ways to reduce your usage to keep your deployment stable.
Reducing build frequency reduces the concurrency of request rates.
Cleanup policies reduce your component count by eliminating components no longer needed.
See Cleanup Policies
Separate your build infrastructure from your distribution method using component replication.