Slow response when invoking the Edit Metadata page

18 views
Skip to first unread message

edwin law

unread,
Aug 12, 2026, 10:02:30 PM (2 days ago) Aug 12
to Dataverse Users Community
Hi everyone

We notice that the response time to enter the Edit Metadata page is longer (around 11 seconds) than before. Below are some facts about our on-premises Dataverse installation:

- version: 6.8 build 1994-92d1ec8
- No. of folders/dataverses: 71
- No. of datasets: 87
- Solr version: 9.8.0
- PostgreSQL version: 16.11
- RHEL version: 9.7
- Payara Glassfish: 6.2025.3
- CPU: 2.8 GHz vCPU x 4
- RAM: 16 GB

Is the response time (11 seconds) much longer than normal/average Dataverse platforms? If so, any tips for identifying the bottlenecks and applying remedial actions? Many thanks!

Loading Edit Metadata Page - Slow response.png

Best regards
Edwin

Philip Durbin

unread,
Aug 13, 2026, 9:57:53 AM (16 hours ago) Aug 13
to dataverse...@googlegroups.com
Hi Edwin,

I wish I had a quick fix for you. No, that 11 seconds doesn't sound normal to me. It sounds quite slow.

In development I've used tools like VisualVM to try to figure out where the bottleneck is. From what I understand, it's possible to use profilers like this in production as well. We have a related page here but nothing about profilers: https://guides.dataverse.org/en/6.11/admin/monitoring.html

Another stat I'm curious about is the amount and percentage of heap space allocated to the JVM. Don and Leonid know a lot about this. Let me at least post some links to some relevant discussions:


I hope this helps!

Phil

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dataverse-community/1dee7fbe-09de-4c04-b6e8-88b8ffdcb2d2n%40googlegroups.com.


--

edwin law

unread,
Aug 13, 2026, 11:32:53 PM (2 hours ago) Aug 13
to Dataverse Users Community
Hi Philip

Thanks so much for your tips. Our current allocations of RAM to the JVM are as follows:

        <jvm-options>-Xmx7992m</jvm-options>
        <jvm-options>-Xms7992m</jvm-options>

Moreover, we have added the following Solr settings according to Dataverse's suggestions (https://guides.dataverse.org/en/latest/developers/performance.html#performance

      <system-property name="dataverse.feature.avoid-expensive-solr-join" value="true"></system-property>
      <system-property name="dataverse.feature.add-publicobject-solr-field" value="true"></system-property>
      <system-property name="dataverse.feature.reduce-solr-deletes" value="true"></system-property>

The hardware usage seems normal:
TopDVMetrics20260814.png

Is the problem related to our customized metadata blocks (the Citation metadata block and the other specific to our institution)? Thanks.

Best regards
Edwin

Reply all
Reply to author
Forward
0 new messages