I’m going to have to try to do that, somehow. But, understand:
I have 29 users
I use gitlab-sso
I’m installed in the gitlab omnibus
So, I have to figure out how to install this thing outside of gitlab. I have to figure out how to migrate my users to standalone users. This is an extreme inconvenience, with zero guidance on how to accomplish this.
It’s very strange to consider ldap or pam authentication to be an enterprise feature. This is not new technology or some tricky implementation.
In this entire thread, there is still no clarity on how the 10K message limit works. Per channel? Per month? Per user?
So, what am I supposed to do? None of the available solutions fit.
In the meantime, gitlab hasn’t included version 10.11, and the apps are complaining every time they’re launched. It would be nice if they could make the apps less annoying during this weird transition time.
Long-standing enterprise features such as LDAP and SAML have always required a paid edition.
That is incorrect. I do LDAP authentication via gitlab. It would be very helpful if MM could provide a migration path, if only to disassociate users with that authentication method.
the 10K limit is enforced to ALL messages over all channels/users. As soon as more than 10K messages are send, an announcement banner is displayed, notifying that only the last 10K message can be searched and displayed.
Thanks for the suggestion. But I don’t think it’s that useful if I already have gitlab running, doing the gitlab-sso. The problem is that MM will be getting rid of that authentication method, so using a substitute that speaks the same language isn’t likely to get me very far.
omnibus is deprecated? What’s running inside of the official docker container?
Any upgrade forces you into the paid edition now, or it tries to. You have to specifically manually upgrade now to avoid this. Personally I’ve just pinned the version onto the LTS release and am awaiting someone to inevitably fork this out of the hands of the corporate-minded people whom are mishandling it and don’t care to address our complaints, to speak frankly. Given they’re announcing breaking changes but can’t be bothered to address this in any meaningful way, they’ve told me all I need to know about their intentions moving forward.
I don’t understand the decision of the company. I understand the desire to make more money. But the desire for the bad PR is strange.
We are living in the age of AI. Shitty AI but AI.
Me and Mr/Mrs AI took for a proof of concept the following amount of time:
1h Setting up a docker test setup
2h Building an automated build system for docker and non-docker build
2h Deleting the server/enterprise folder and dealing with the missing dependency fallout.
6h Adding a new SSO OIDC mechanism from ground up, intentionally not using the gitlab stuff. i.e. isolating the new SSO OIDC from the changing Mattermost code as best as possible.
0.5h Remove the user limit.
3h Making a new logo and put it in.
8h (+ not done yet) De-branding (i.e. renaming it to something else) and removing the dead enterprise UI elements.
Lets say it took a weekend for me and the AI (only with free tokens) to make the code fully open source + working + limit break + SSO OIDC. And I am not able to program in golang or react. For a real full stack programmer, it will take less than we all took to write the posts in this thread.
I am now waiting for my law department to check on that proof of concept. If I get the green light, then I will de-brand the remaining 2.5 pages and then we can use it as a replacement. If I can do it, everybody can do that.
The new pricing options imposed by Mattermost have put our non-profit organization in a difficult situation. We currently have about 600 users, and this will grow to eventually exceed 1000.
We do not have the resources (like most non-profits) to pay the commercial licensing fees.
Even if we pruned inactive users, it would be nearly impossible to stay under the 250 user cap for the Teams edition.
Even though we are willing to pay for a non-profit license, we will eventually exceed the 1000 user cap imposed by Mattermost, and there is no guarantee that Mattermost won’t reduce this cap in the future (their announcement seems to imply that further user reductions should be expected).
Switch to an alternative platform, of which the only contender is Matrix.
As a stopgap solution, we will likely user Mostlymatter and make a $250 donation to Framasoft. However, if a community supported fork doesn’t materialize that implements more than just the user cap patch, we will likely switch to Matrix in the future. Though Matrix is not ideal, the public sector in many European countries are switching to Matrix, and this will act as a counterbalance to any enshittification that Element might impose.
I worry that the decisions of Mattermost will only bring them short-term profit at the expense of degrading or even killing the project in the long term.
But the desire for the bad PR is strange.