I see too many issues that point to basic functionality not working in the 6.5 releases. Is there a release the team suggests to use that has the least amount of regression bugs? I can't go off of the regression label in github because so many more issues should be labeled as regressions but aren't labeled as such.
We wanted to upgrade from a 5.6 release to a 6.5 release because of some new Elasticsearch features which unfortunately means we have to upgrade Kibana to one of the 6.5 releases that have too many bugs that make it an unusable product.
From observation, I'm starting to wonder if 6.5 was rushed to release. Basic functionality like filtering on some of Kibana's most used visualizations does not work. There has also been 4 different releases for 6.5 in the last month.
Let's look at some code
6.5:
Master:
The code linked above shows that the 6.5 release is a mix of API choices where some of the visualizations handle events one way and the others a different way. I'm guessing because of the continual use of some of the legacy response handlers in these visualizations. It's not surprising that on master it can't be reproduced when the code linked above for master has changed to another different design on how to handle the filter event for visualizations. This same change is on 6.6 branch.
The question then remains should they fix issues like this with in a 6.5 release with the design it was released with or the new design on master. Keep in mind the 6.5 code I referenced has changes to that method that is only 2 months old and presumably was tested/reviewed/worked at some point before release.
I know some of my frustration of this release is coming through in this message but sometimes you gotta let them know otherwise things never change and only get worse. I also know releases without issues is impossible but these issues are rather silly to have fallen through. It feels like the pace at which Kibana is rolling changes out needs to slow down. Otherwise they're looking at reaching 6.5.10 at the end of January. Users are then left wondering what new bugs & loss of functionality will they have to deal with in a newer version just to have a fix for a bug they're currently dealing with on an older version.
I had issues when I upgraded from 4.3. to 5.6 but ones that the team didn't seem interesting in fixing that I had to fix. To be fair it was an edge case that not many people probably utilized that feature. However, this time it's just an unusable product that no one should deliver, so it's definitely getting worse hence the rant.