Repository navigation
Bug: UI health state doesn't reflect the latest state #5450
Description
Activity
I cannot replicate this locally with the sample servlet app, though.
I wonder if the UI has somehow some caching issues or what could the cause be...Hi.
Indeed it seems like there is an issue there, but it's more related generally with the refresh of the UI itself and not with the change of state from UP to DOWN or whatever.
See for example:
The state was not updated since yesterday in the UI.
Hey @SteKoe, @ulischulte, @hzpz.
Can we have the provided fix reviewed and released as soon as possible, please? Tnx- added 10 commits that reference this issue
on Jun 17, 2026 7 remaining items
- added 4 commits that reference this issue
on Jun 19, 2026 - added a commit that references this issue
on Jun 19, 2026 @SteKoe while the in-memory event store implementation works just fine with the work that was done here (by setting the property exposed in #5483 to
fullexplicitly), with Hazelcast I'm having again this issue where the status shown in the UI is stale, even while keeping the same configuration said previously.It's not only stale/old in sense of content, but even for the status itself.
I've an health indicator which is reported asATWORKsince 1 day, but if I call directly the health endpoint myself the application reports allUP.As a matter of fact, it seems like SBA is not even pulling for updates anymore (at least, according to the journal view)
since the last entry is from yesterday.
Do we want to use this same ticket to investigate this or should I open a new one?
I had to introduce the Hazelcast event store because I enabled notifications.
I don't know if there is some other way in SBA to keep the data consistent across multiple instances.@SteKoe while the in-memory event store implementation works just fine with the work that was done here (by setting the property exposed in #5483 to
fullexplicitly), with Hazelcast I'm having again this issue where the status shown in the UI is stale, even while keeping the same configuration said previously.It's not only stale/old in sense of content, but even for the status itself. I've an health indicator which is reported as
ATWORKsince 1 day, but if I call directly the health endpoint myself the application reports allUP.As a matter of fact, it seems like SBA is not even pulling for updates anymore (at least, according to the journal view)
since the last entry is from yesterday.
Do we want to use this same ticket to investigate this or should I open a new one?
I had to introduce the Hazelcast event store because I enabled notifications. I don't know if there is some other way in SBA to keep the data consistent across multiple instances.
I've a solution for this. I'll submit a PR in the next days.
It was closed due to a merged PR.
Reacted by Cosimo Damiano Prete
Spring Boot Admin Server information
Version: 4.1.0
Spring Boot version: 4.1.0
Configured Security: None
Webflux or Servlet application: WebFlux
Client information
Spring Boot versions: 4.0.x
Used discovery mechanism: Eureka
Webflux or Servlet application: Servlet
Description
Hello.
It seems like the fix provided in #5286 is now preventing the UI to show the latest health state correctly and, instead, some stale value is getting shown.
See, for example, the picture below:
where the latest state is the one in the JSON on the right and it should, therefore, be DOWN, but the health indicator in the UI is still shown as UP.
Weird enough, the journal view doesn't show additional
STATUS_UPDATEDevents either apart maybe the 1st one, whose payload matches the UI shown in the UI.It's like the newer state is never produced by the SBA server, like if the StatusUpdater is stopped.
If it's not the case, then somehow the new states are not getting propagated, either in the back-end or in the UI.