Hacker News Reader

Comment by siliconc0w | Hacker News Reader

Asking users to run mongo, es, neo4j, and redis is a tall order. Mattermost just needs (AFAIK) a relational database.

Also, less relevant but the Gitlab UI has far too much wasted whitespace. Like 85% of the page is navigation. I like open source competitors to incumbents but maybe just copy Github here - no one is gonna say anything.

View on HN

Hi, co-founder of Gitter here.

Gitter is not intended to be a replacement for Mattermost, Slack or other team collaboration tools. We see Gitter as a community instead.

As such, we're not expecting to see a huge uptake of on-site installations, so the list of required services (es, neo etc) is big compared to other products focused on on-prem.

We're hoping that our users will contribute to the main site, Gitter.im.

Obviously, we're also totally happy with users running their own Gitter installations but, while we would like it to be easy, ease-of-installation of a production instance is not a goal currently.

> we would like it to be easy, ease-of-installation of a production instance is not a goal currently.

ok but doesn't it behoove you anyway to have e.g. dockerfile and/or ansible roles for dev/staging environments?

They probably do have some sort of automation, but it's likely specific to their environment.

We use Ansible / packer & terraform. Open sourcing this repo will be fairly complicated but it may make more sense for us to publish a k8s helm chart or more complete, production-ready docker-compose.yml

you are already using docker-compose.yml. you should use definitely do a production ready docker swarm deployment. you only will need to do incremental changes to your docker-compose

Make a merge request.

We have a docker-compose.yml for development environments and I'm currently in the process of simplifying this[1] so that Docker for Mac/Windows will be able to spin up an environment with little effort.

We use ansible for provisioning beta/staging and production. We have yet to open source the ansible repository but, since we're switching to GitLab CI/CD, the deployment process will soon be publicly accessible - even for production.

1: https://gitlab.com/gitlab-org/gitter/webapp/merge_requests/1... so that

What would be the hardware requirements to run gitter for a couple of hundred accounts?

We haven't done much benchmarking at that end of the scale, but I can't think of any reason why you wouldn't be able to run it on a single server.

Sure, but is it a 5USD/month DO instance or a 640USD/month one?

The easiest route for installations would be to offer a docker image and docker-compose file to set up the multiple services.

We currently use docker-compose for the development environment. I'm currently simplifying this to allow developers to get started with minimal effort:

https://gitlab.com/gitlab-org/gitter/webapp/blob/feature/mon...

Going forward, I agree that we could offer a more production-like docker-compose setup and/or Kubernetes Helm Charts.

This is actually the first real project I've come across that uses neo4j - thank you for pointing that out.

Given the requirements, I actually think this project looks rather clean between the docker-compose-file and the npm dependencies.

I very much doubt it can compete with Mattermost or Rocket.chat for self-host for small organizations - but I'm really happy to see this thing open sourced!

I much prefer it to be opened up as soon as possible, and then flicked on later.

If there's one thing I'd like to see (beyond a reduction in what appears to be a default of three? mongodb instances) - it would probably be a set of juju charms for bringing the thing up under lxc/lxd on Ubuntu.

I just revisited the combination of juju+ubuntu+lxd - and while I was skeptical when juju was introduced, it appears to be a rather sane toolkit for automating deployment IMNHO.

For those interested, if you start here:

https://insights.ubuntu.com/2016/06/08/lxd-2-0-lxd-and-juju/

you might think that this thing is insane overkill - but if you start at the beginning of the series[s], you will (or I did) get a warm fuzzy feeling about how much fun it is to play with lxd on top of zfs. I suppose some people think docker is just as fun (and warm and fuzzy) - but I don't know. LXD does feel like it combines the good parts of linux with the good parts of bsd jails.

[s] https://insights.ubuntu.com/2016/03/14/the-lxd-2-0-story-pro...

(Gitter Co-founder here)

Unfortunately our neo4j setup as it stands really doesn't scale. At an application-level we're built error-handling to gracefully handle the frequent outages we experience running it.

These are probably our fault, rather than a failing in the product, but I've wanted to replace the current neo4j-based suggestion algo with a new one that uses batching/clustering for a while. As soon as I get a chance, I would like to remove the dependency we have on neo4j.

You might want to reach out to Al Baker or Kendall Clark at Stardog. Their product is more RDF centric than for generic graphs like Neo4j, but is in my opinion a much better product. They have a community version, and have open sourced several pieces, but it is commercial. https://www.stardog.com/docs/

If you want pure Open Source then Titan might be a good fit for Gitter. http://titan.thinkaurelius.com/

The big problem with the way we use neo4j is around huge rooms with ~100K users. The number of possible rooms that neo4j has to traverse (even a shallow traversal) for a user in a big room is huge.

Thanks for your recommendations. I'll definitely check them out.

Those huge rooms really don't add anything to recommendations (like everyone buys milk), if almost everyone is in the room you can also just leave it off (the same is true for rooms with very few users).

That's why in recommendations, I usually filter them out b degree (which is a constant value read).

MATCH (r:Room) WHERE size( (r)<-[:MEMBER]-() ) < 5000) ...

You're also not using directions in your queries, which will also help.

I looked at your other queries, they can also be improved easily, if you want to I can help you do that. Just need some sample data or a read-only access to the db. Just ping me at michael@neo4j.com.

It looks like you're still using Neo4j 2.3. Have you tried using newer versions of Neo4j? There have been some pretty substantial performance improvements.

Additionally, Neo4j does support clustering with their Enterprise edition. It also has much better tooling (better metrics/logging, backups, etc.). It is AGPL licensed, so there's no reason not to use it really.

you should really look at Janus (with Cassandra as the backing datastore). http://janusgraph.org/

Since it uses Gremlin, you should not have that big a problem in porting.

Why not get in touch with us and let us help you instead of dropping us?

I'm totally open to that.

I'm not going to be changing anything around the suggestions any time soon, but who should I get in touch with when I do?

Email me: max.demarzi@neotechnology.com

> Also, less relevant but the Gitlab UI has far too much wasted whitespace.

And why do I always need to scroll down to see the project README?

You mean exactly like with Github?

Solution: https://gitlab.com/gitlab-org/gitter/webapp/blob/master/READ...

But now I'm having problems with the length of the URL :)

PS: yes, same problem with github

There is a small hack on Gitlab for that actually: right under the textbox with the clone URI there is a link to the README.

Files (209 MB) / Commits (44,194) / ... / Readme / MIT License / Contribution guide / ...

Personally, I'm finding myself far more often interested in project's main files/dirs update dates that project's readme.

Perhaps a two column layout for big wide screens would be best. That would cater for both use cases.

> Asking users to run mongo, es, neo4j, and redis is a tall order.

Not really. Why? Reddit is opensource but it'd damn hard to setup correctly and start your own instance.

Just because Reddit does it rather badly does not mean that other OSS projects shouldn't try to do better.

>Just because Reddit does it rather badly

What about other OSS project? It's also not easy to make Linux from scratch. Being OSS doesn't mean it will be single click to run your instance.

Linux doesn't really have to since 99% of users use distributions, not LFS.

And there's a difference between gratuitous (or almost gratuitous) dependencies and must have ones. From what I read in the other comments Neo4J doesn't feel like a must-have for this project.

Plus, to be honest, Mongo, Neo4J, Redis AND Elastic Search? That's got to be overkill :)

The reddit repo doesn't have any commits since March. If that's their actual codebase, what are all their engineers doing?

It's not the actual codebase, it's kind of a branch that has to be manually updated (by cherry-picking across commits from the "real" one). It used to be kept in pretty close sync with the code live on the site, but it seems like they don't care about updating it any more. There are things that have been live on the site for close to a year now that aren't in the open-source repo.

That's roughly what I had guessed. Thanks for clarifying.

Hiya,

Sorry to hear you're displeased with our UI. We'd love to hear any suggestions you might have. We're always looking to improve as much as we can. You can open an issue in https://gitlab.com/gitlab-org/gitlab-ce/issues

Matrix is also quite good on this. Homeserver and PSQL (and maybe nginx), done.

Gitlab UI is the worst. I hope they do less features and just fix their god-awful UI for once!

Interesting (and rude). I think it's much better than any other group chat app. Especially for groups organised around code projects.

I was talking about GitLab, not Gitter.

Yes it's rude, but I want to love GitLab - it's amazing. However I just can't bring myself over with this UI. It's just too bad and too unintuitive. Also slow.

Oh, sorry I missed that you said GitLab. It's less rude if you're not commenting on the open sourcing of a project to just say negative stuff about it :)

I have to use GitLab for work and I find the UI much harder to use than GitHub, so I agree there. But, it's open source and we can all contribute, and in particular the best place to make your feelings know about the UI is constructively in the issue tracker.

Here's the main issue sparked from the last HN discussion about GitLab UI: https://gitlab.com/gitlab-org/gitlab-ce/issues/25752

From my own experience, Gitlab is fast for as long as u have sufficient RAM available. It becomes slow when the SWAP is hitted.

> Also slow.

Only on gitlab.com, on prem here it flies.

Same here. Extremely fast on a 4GB VM.

This type of feedback is almost completely useless for a project. "the worst", "too bad", and "unintuitive" are not actionable so if you really want to help a project then describe concrete issues.

I'm not an UI/UX expert but a back-end developer. I cannot give you more feedback then this.

"I'm not an UI/UX expert but a back-end developer."

This actually makes you the perfect candidate for giving us good feedback. You may not be able to articulate exactly why UI elements feel off or don't work for you, but you know they could be better. If you are interested, you can join our UX Research Panel to help us pinpoint areas we can improve. We want you to love GitLab too :)

https://about.gitlab.com/researchpanel/

Hiya,

Sorry to hear that you're displeased with our UI. We'd love to hear any specific feedback you might have. You can also open an issue in https://gitlab.com/gitlab-org/gitlab-ce/issues

We're always looking to improve

A relational database is overkill. My BBCode-enabled chatroom runs off txt files.

Mattermost isn't exactly a simple bbcode-enabled chatroom though. They have support for multiple projects, an API etc. Achieving this with just text files would be quite the headache.

They also have horrible android client and you have to pay for mobile notifications.

I ended up using matrix/riot instead. Much better clients and grows nicely.

Hi omnimus, thanks for the mention.

Mattermost has a brand new, open source Android client in beta written in React Native used by thousands of testers: https://play.google.com/store/apps/details?id=com.mattermost...

It's greatly improved from the original version and official release is this month.

Mattermost mobile notifications aren't a commercial feature, they are just off by default due to privacy (push notifications appear on a mobile device when the device is locked).

Go to System Console > Notifications > Mobile Push > Send Push Notifications to enable.

You can also deploy your own Mattermost Push Notification Server (https://github.com/mattermost/mattermost-push-proxy) and compile your own Mattermost Android and iOS apps (https://github.com/mattermost/mattermost-mobile) for total control.

Sadly the Android's react version seems to lack any share features - the app doesn't appear in any 'share to' dialogues, nor can I share pictures or such from the app itself (For example using rocket.chat I can long press on an image and I will be able to share the picture to other apps). The issue appears to only have around 100 combined votes on uservoice (https://mattermost.uservoice.com/forums/306457-general/sugge... + https://mattermost.uservoice.com/forums/306457-general/sugge...) so I'm not holding my breath for this to be implemented anytime soon, but it would be a really useful feature to have for when you need to move graphical assets around and you want to take a quick look on your phone for example.

> Asking users to run mongo, es, neo4j, and redis is a tall order

They do provide a docker-compose.yml, which you can use in Docker Swarm, ECS, etc to stand up the entire stack.

Reply on news.ycombinator.com