Agreed, I tried self-hosting it a couple of month ago and it was impossible. I spent the day on it but the setup process was broken because of a recent change which was made for their cloud offering. Managing both a codebase both adapted to a cloud deployment with a huge amount of users and to a self-hosted way small deployment is very hard and requires a lot of resources. It's hard to justify investing this much time and money in making it work well for a self-hosted setup, and it seems like they stopped doing so.

It's still great to make the code open, but it's not usable anymore for a self-hosted setup.

It's also worth noting that in 2023 they abandoned their Kubernetes support which was relied upon by a full 3.5% of their users: https://posthog.com/blog/sunsetting-helm-support-posthog

In their rationale for this:

> We also learned that the tools to do that automation just don't exist. We kept finding new failure modes. When onboarding a new customer we would have to vet their engineering team for Kubernetes experience so that we'd be confident they could help us debug issues in their PostHog deploy. Folks that didn't have infra experience would often be able to get something set up, only to get stuck when something went wrong.

I empathize that this is a sane choice for PostHog to make as a business. But - if you can't deploy and dogfood your changes, are you truly able to maintain a fork with customizations? And if you can't use your own changes, is the software open-source, or source-available?

Perhaps the punchline is that any scalable & performant web analytics platform must necessarily be a distributed system of ingestion and storage services, and that complexity is like oil and water with the classic "you should be able to swap out the dependencies on your systems with ones you fork" open-source ethos.

PostHog had an opportunity to break this trend, to innovate and invest in those automations they correctly said didn't exist - and I was cheering them on. I've been saddened to see them move in the opposite direction.

It seems like an odd thing to run locally with so many dependencies.

They want you to buy their hosted service, that's where the convenience is sold. If they give you a one liner script you can paste in or a docker compose that does everything from scratch they cannot sell their hosted services.

we went from batteries not included to BYOAi

Would love for you to try Amplitude. We've put a lot of work into making sure the core is usable. We've also started to fix a lot of the most common complaints about our pricing.

They've been using AI to shove a lot of AI into their product and trying to force everyone to use AI. I really don't understand the why of any of it. The product was working great for what it needs to do. I don't need AI to make guesses about data for me and I especially do not want _yet another product_ trying to write features in my codebase (which is their latest push).

PostHog has always been open source. I'm not sure why this has been shared on HN at all.

I'm normally not a fan of AI being shoved into products but I love it in PostHog. I can ask it a product question, ask it to slice and dice data, etc and it does a really good job of sifting through my events, doing analysis, building dashboards, etc. It's normally boring tedious work to do that stuff in PostHog so I'm impressed and glad it works.

PostHog has always been open core and source-available. What's new is that they now publish a GitHub repo containing just the open source parts.

Has it? This is from that repo, "PostHog FOSS is a read-only mirror of PostHog, with all proprietary code removed." Seems they have some proprietary code in the non-open-source version.

I thought the same thing. What changed, if anything?

If you are on react/next.js, defer the client side initialization until after your app has painted. PostHog (especially with their session recording feature) likes to initialize a little before the rest of your app loads through either their context provider or instrumentation.ts (if you are next.js). It's generally insignificant except if you are building a B2C web app where the extra 100ms makes a difference in retention.

On the server side, queue all analytics call and run them after the main request completes (on next.js it would be within an after() function on the server side).

You can paste this comment into Claude and it should handle the refactor just fine.

Note with the changes in analytics scheduling, you will lose out on real time analytics in favor of better time to first load. So keep the trade-off in mind.

Spent a day trying to get it running 2-3 weeks back. Gave up after a day.

How is this relevant ? We're talking about open source analytics, and this looks like a shameless plus

Open* is not Open

wait i don’t see this happening!

Yeah weird, bc I just checked and it isn't oss. Only the SDKs are OSS, which...doesn't help much

Why would you apply to a job for a company you don't even know what they do?

I used their AI chat last night and I was impressed with the product implementation. I was able to use it to make quick sense of data and even generate prompts to solve problems locally.

IDK what the prior AI behaviors have been, but for me, what they have now is an ~idealized version of AI product.

It is basically what I'd do if they didn't offer it but not as well: export data, import their docs in md., import some industry best practices in analytics into context etc.

Except its all right there. Not sure of the economics of it, as it was on a free account and some reasonably good model was powering the discussion. But it was about as engaged as I've been with an analytics tool.

Exactly, I was thinking this too!

money?

founder here, what specifically do you use posthog for that you now find hard to find?

Couple things for me: - I missed the default dashboard being my homepage and not having to navigate to it via an extra step. I just noticed that i can change that default, but personally found that a bit annoying largely because I felt lost.

Also i don't like how "People" is not buried in data now, pretty sure this was just in the left hand bar. I can "star" it, but its still nested in a wierd UI drawer. I wish it was pinned by default, or a way I could re-pin it.

A message of some kind (You may have already done this) to accept the new defaults or keep the some of the original defaults would of been cool.

On a broader note, I loved the simplicity of what you guys made, and it feels like the product has too many moving pieces available to me. I build my own product as well and I get the rub here, you want to support so many more use cases and re-organize things to do so , but in the end it feels bloated. I love the idea of using AI and in some cases when i needed to mine my data some it was super nice, though most of the time from a founder perspective I'm looking at largely the same data/metrics over and over, and don't need to be creating new graphs every day. So the launchpad /w chatbox as the primary thing as a default feels a bit lost on me.

All that said, I still looooove your product and would still use it 1100 times over GA any day. <3

It's all the AI bloat for me. It keeps getting shoved in my face and they _really_ want me to use it. But I neither want nor need it. The product itself was so good that it doesn't need this whole guessing-machine-layer on top of it.

I also agree that the nav, structure, and defaults changing keep being un-intuitive and are making it harder to find the things that I use and that matter to me and my orgs.

I dont think is something in specific but I used to love the simplicity of PostHog and that's why I pushed my team moving to it from Google Analytics.

That has changed though. Last time I used PH, I found myself not finding the things I needed so ended up connecting the mcp and asking claude to change things for me.

MCP was a nice addition, so thanks for that. The problem of mcp is that it prevents product discoverability. You are adding many things lately and since I'm using the MCP for my "part of the job" I will never visit your site anymore and see what's new.

I've notice that as well. If you don't know how to do something having an agent with tools is great (in general, not just PH). But, its also quite slow and imprecise. I think that having the agent act more as both a teacher and "doer" is best.