What is the reason Elastic introduce and release different version at fast pace?
it is very hard to keep up with it.
And if you fall too much behind version then upgrade process gets harder and harder.
for example I was stuck on 7.17 for long time and finally upgraded to 8.5.2 and they already on 8.6 in less then month.
If I'm not wrong they are releasing a new version basically every month and version 8 was released almost 1 year ago, it is a steady pace bringing new features and fixing issues.
But unless it is a critical security fix, a bug fix that impacts you or a new feature that you want to use, I do not see any reason to update so frequently.
For example, I need to set a maintenance window to update my cluster, so we schedule some around the year and update only two or three times a year, currently I'm on
8.5.2 and the next update will probably get
8.7.1 as we avoid updating to
It took me while to update from 7.17 to 8.5.3 ( because it requires lot of testing and validation) due to major version change.
It's mostly closer to every 2 months, sometimes a little longer. It's pretty consistently been 6 per calendar year for a while now:
$ for i in $(git tag | grep -e 'v.[0-9]\+.0$'); do git log -n1 --pretty='%cd %d' --date=iso $i; done | sort
2019-04-05 17:42:58 -0400 (tag: v7.0.0)
2019-05-15 19:04:05 -0400 (tag: v7.1.0)
2019-06-20 15:30:13 +0300 (tag: v7.2.0)
2019-07-24 11:38:07 -0400 (tag: v7.3.0)
2019-09-27 09:04:09 +0300 (tag: v7.4.0)
2019-11-25 16:23:05 -0500 (tag: v7.5.0)
2020-02-05 14:00:43 -0800 (tag: v7.6.0)
2020-05-08 16:11:44 -0700 (tag: v7.7.0)
2020-06-12 16:39:07 -0600 (tag: v7.8.0)
2020-08-11 12:42:47 -0400 (tag: v7.9.0)
2020-11-05 16:29:12 +0100 (tag: v7.10.0)
2021-02-08 11:14:28 -0800 (tag: v7.11.0)
2021-03-17 21:48:56 -0400 (tag: v7.12.0)
2021-05-19 14:15:03 -0600 (tag: v7.13.0)
2021-07-29 15:18:55 -0400 (tag: v7.14.0)
2021-09-15 15:26:58 -0400 (tag: v7.15.0)
2021-12-02 08:04:26 -0500 (tag: v7.16.0)
2022-01-27 17:05:30 -0500 (tag: v7.17.0)
2022-01-31 17:41:04 +0200 (tag: v8.0.0-rc2, tag: v8.0.0)
2022-03-03 07:57:03 -0500 (tag: v8.1.0)
2022-04-20 12:08:52 +0200 (tag: v8.2.0)
2022-06-22 08:53:05 -0400 (tag: v8.3.0)
2022-08-19 08:16:40 +0930 (tag: v8.4.0)
2022-10-24 16:58:18 +0200 (tag: v8.5.0)
2023-01-03 21:44:10 -0500 (tag: v8.6.0)
There's no need to update each time, but the option is there if you want to. It seems better that way than holding releasable features back for no particularly good reason.
most time I feel needs to upgrade is due to FOMO (fear of missing out)
But yes I know if you left behind too many version then it will be hard to upgrade as there might be some breaking changes which you might miss it and upgrade fails. which happened at this upgrade cycle
As far as possible we avoid breaking changes in minor releases, and call them out in the release notes if they're unavoidable, so upgrading from one 8.x release to the next should be safe. Major releases are more work but we try and help as much as possible, and there's only been two of them in the last 4 years.
I can understand the pace, I can't understand the reasons I have started with Elastic Stack 8.2 not so far ago and now is 8.6. The problem is that not always the upgrade runs smoothly. I have upgraded the Elastic from 8.4 directly to 8.6 and now my Kibana doesn't connect to Elasticsearch. I have to check the config files now (most of it have forgotten). I get the message "Kibana server not ready yet". Anyone having the same problem? Or debugging is what I have to do, again
Minor upgrades are supposed to be smooth, so that's definitely not supposed to happen. Please open a bug report.
I also had some minor issue for each upgrade.
open up new thread to solve this as people might not look this due to different title
Open up a bug report.... Sorry for my sarcastic reaction, but Elastic has "quite a bit" of bugs and problems after (almost) every update. If I had to make bug reports for every issue we have in Elastic I'd need a full time job for that.
Besides of bringing out new fancy features every update, imho Elastic should also spend more focus on making sure their stuff works as advertised. Also the rate at which stuff which worked perfectly for years is getting deprecated, is very hard to keep up with.
Tell me about it deprecating Multi disk (multi data path) on a node, and everyone is hammring them still they are going with it.
Hi @willemdh, I hear your frustration but I'm not sure what to do to help. We put substantial effort into making sure that things continue to work across upgrades and set a very high bar for accepting breaking changes. We have accepted just two breaking changes in the 8.x series so far, both bug fixes with limited impact:
We take feedback and bug reports seriously, and also sometimes invest significant time into improvements in our QA process itself in cases where there's a pattern of feedback showing that it needs improvement. But it's hard to act on general statements like "quite a bit of bugs and problems after (almost) every update": without specific information we don't even know where we are going wrong.
@DavidTurner Although it's tempting to start listing up some issue we have and have been encountering lately, it would lead into further discussions and I currently don't have the time for that.
Just to clarify => I love Elastic and all the things it can do. Really. I've been using it for a long time and it keeps amazing me. I have seen no other products which can do the things Elastic can do.
Hopefully in a few months, things calm down at work and maybe I find the time to try explain them to you (or Elastic support) in the hope for a solution (for some of them)
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.