5 ms·
Show HN: I built a self-hosted error tracker in Rails
This project is inspired by 37signals’ ONCE idea. I replicated the whole process and have already sold a few copies (the testimonials are real).
- keyle 11mo agoIt looks like it was built on macOS and not Omarchy, is that even allowed these days? /s
- kyrylo 11mo agoHaha, true. I used to rock Arch Linux back in 2012–2015, so no need for Omarchy for me
- johneu88 11mo agoNice. I’m curious about what kind of customers are buying your product and where did you find them. It’s difficult to compete with someone like Sentry, and even though you have a better pay model, you don’t have the reputation, so I am rather curious about your customers and why they prefer your product. Are they indie devs looking for a cheaper solution?
- kyrylo 11mo agoThanks! I’m not really competing with Sentry itself. I’m competing with self-hosted Sentry, which is notoriously hard to install and maintain (and has steeper hardware requirements). I'd say my customers prefer my product because: - They want self-hosting without the maintenance burden. - They work in regulated or internal networks. - They’re tired of subscription pricing. - I build it in public and post regular updates on my social media. - They value direct support from the creator. P.S. I’ve personally worked for a Sentry competitor, so I know the pain points firsthand.
- bberenberg 11mo agoI think this is a great idea with the wrong pricing model. Look at all the one off payment products that involve code, they're all dead. Just charge a lower but recurring price so I can be sure that you make enough money that you keep working on it. $20/month flat price to keep the license working and source available if you shut down. If people like it and want the barebones Sentry then charge them $40 a month to provide code and host it. Wishing you the best of luck.
- kyrylo 11mo ago> Look at all the one off payment products that involve code, they're all dead. Could you share some examples? I can’t think of any off the top of my head. -- I really went all-in with the ONCE philosophy because it resonated with me deeply. It felt more like a passion project than cold business strategy.
- bberenberg 11mo agoI think all of the boilerplate projects you can find. Are ONCE projects getting updates? We will find a year or two? Your model is a subscription, we just don’t get to know when you decide to have a new major version and plan pricing / spend as a result.
- kyrylo 11mo agoONCE projects do get occasional updates. I don’t use boilerplate projects much, so I can’t speak to them. My model isn’t a subscription. Think about it like buying rice. You might buy it every week, but that doesn’t mean you’re “subscribed” to rice. Even if I release a new major version, you’re free not to update. And if it’s a major version, it’s fair to expect it to be paid. After all, major updates usually bring significant improvements. For example, if you played the original DOOM, you had to pay for DOOM 2 too, even though they run on the same engine.
- pmontra 11mo agoThere are so many people buying rice that's not hard for rice producers to forecast how many thousands of tons of rice they have to grow each year. And not many people look at how much they spend for rice year long. For the single box, yes, they do notice. So it's not a subscription but it looks like it is, at least from the point of view of the seller.
- bberenberg 11mo agoYes but I don't rely on rice I bought a year ago or DOOM as a core component of my business. Trying to work around a business model (subscription saas) requires that you understand what people are buying, and often, especially with you vendors, that is a financial alignment between the two.
- krystofee 11mo agoTheir docs show throughput limits (e.g., 4 CPU = 60 errors/sec), but what happens during error spikes? If my app crashes and blasts hundreds of errors in seconds, does Telebugs have built-in rate limiting or backpressure? Or do I need to overprovision hardware/implement throttling myself? With SaaS tools, spike protection is their problem. With self-hosted, I’m worried about overwhelming my own infrastructure without adding complexity. Anyone running this in production?
- kyrylo 11mo agoHey, Telebugs creator here. Great questions! Right now, Telebugs doesn’t have built-in throttling, so during error spikes, you’d either need to handle it manually or overprovision. I do plan to add throttling in the future, similar to what Sentry does, to protect your infrastructure automatically. Curious: for those running self-hosted error trackers in production, how do you currently handle sudden error spikes? Any clever tricks or patterns you swear by?
- johneu88 11mo agoThe company I work for runs self hosted sentry. Sentry has something that tells you that events are being dropped due to pressure. I think every engineer in the company knows that this is happening but no one fixes it because no one has the time to look into it.
- kyrylo 11mo agoThanks for your answer! Would you mind sharing your error volume? I’m also curious, how often do dropped events happen, and how does it impact your workflow? Any workarounds you’ve tried, or features you wish were available? This will help me make sure the feature is implemented in a way that’s actually useful.
- IshKebab 11mo agoThat seems like insanely low throughput. What takes it so long?
- 11mo ago
- dbunskoek 11mo agoIn this Sentry alternatives space there’s also Bugsink (https://www.bugsink.com/ https://www.bugsink.com/). Can heartily recommend it.
- BirAdam 11mo agoYeah. Free, works, and saves my employer a ton of money after we started switching over to it.
- euph0ria 11mo agoHow does it compare to GlitchTip which is in the same space?
- kyrylo 11mo agoTelebugs creator here. GlitchTip is solid, but Telebugs is built for devs who want something lighter and faster to self-host. It runs on plain Rails, sets up in about 5 minutes (one command), and stays snappy even on small servers. The UI is modern, minimal, and actively maintained. I keep refining it to stay fast and clean. The biggest difference is in philosophy. GlitchTip was built by an agency. Telebugs is a solo passion project. I’ve worked on error tracking tools professionally before, and built Telebugs to reflect how I wish those tools worked. If you’re curious, here’s a short write-up on why I built it: https://telebugs.com/why https://telebugs.com/why Happy to answer any specific questions!
- nodesocket 11mo agoDoes Telebugs support SQLite3 query performance tracking? E.G. top slowest queries, etc. I don't believe Sentry supports this
- kyrylo 11mo agoNope, there are no APM features in Telebugs: https://telebugs.com/#okay-but-sentry-offers-apm https://telebugs.com/#okay-but-sentry-offers-apm Fun fact: Telebugs itself runs on SQLite3!
- nodesocket 11mo agoWould love to see it added. I’m running a Flask app using raw SQLite3 without a ORM. I believe I could hack up a Span solution in Sentry to query track like: import sqlite3 from sentry_sdk import start_span def execute_query(conn: sqlite3.Connection, sql: str, params: tuple = ()): with start_span(op="db", description=sql) as span: span.set_data("db.system", "sqlite") cursor = conn.execute(sql, params) return cursor
- kyrylo 11mo agoTelebugs isn’t meant to be a full Sentry replacement. It focuses purely on error tracking. The goal is to avoid feature creep and do one thing exceptionally well, much like classic UNIX tools. That said, you’re not the first to ask about APM features, and it’s motivating me to consider building a separate APM product.
- ohr3xooqu9vi 11mo agoThis is cool; I have been watchjng this for a while and intend to use it soon.
- kyrylo 11mo agoThank you!
- riddley 11mo agoIs there a license for FOSS projects?
- kyrylo 11mo agoNot at the moment, but I’m open to it! If you have something in mind, feel free to email me at kyrylo@telebugs.com and we can work something out.
- dewey 11mo agoI know you said inspired, but it still feels a bit wrong to just copy paste the exact same landing page style and replace some words on it (https://once.com/campfire https://once.com/campfire)
- kyrylo 11mo agoYeah, I used it as a starting template. But beyond the layout, the copy and product are completely different. It just helped me get started faster, and I’ve always been open about that. Everything else (including the installer) is fully self-made.
- replwoacause 11mo agoIt's not the exact same at all.
- jlengrand 11mo agoInteresting pricing model, and it seems the pain of Sentry is getting real those days. Many folks need something simpler, that just works. Totally support the idea. The alternative starting with a G has been mentioned a few times already, so I'll also mention that's been operating in the space and that I happily use : https://www.bugsink.com/ https://www.bugsink.com/. It's trying to solve the exact same pain. Disclaimer : I know the owner :), so I may be biased. But generally I like to see more niche alternatives to the massive players in the field
- vanschelven 11mo ago[flagged]
- vanschelven 11mo ago"it can deal"... with 300 events/s easily.
- nasmorn 11mo agoIt works out because sentry has basically gone full solarwinds. Trying to get rid of the small accounts because they only care about the enterprise market anymore
- zeeg 11mo agoClearly this is bait but we have done nothing but be supportive of small accounts. It’s still $29, we still are extremely generous on forgiveness and sponsorships, and we still prioritize the long tail. 70% of our business is smb/mm, and if you knew anything about us you wouldn’t make the statement you said above. If we have done something to show otherwise I would love to see it so we can address it.
- nasmorn 11mo agoI guess I’ve got a lot of resentment for when I suddenly paid for 100k reserved errors suddenly that I didn’t remember asking for. I probably figured out I have control about it at some point and that’s why I decided to not do the Schlepp of changing the error reporting on all these small projects. So while I can actually corroborate that your cheapest plan has not changed my takeaway is that changes to billing just make people unreasonably resentful. At least me.
- newscracker 11mo ago> Wherever you can host something like WordPress, you can host Campfire I’m going to be pedantic here, but this statement is not true. I host a website on a provider that allows WordPress (PHP) along with MySQL, but > System requirements & installation > Campfire is packed as a Docker container image the web host provider does not allow Docker (it runs on BSD). I’d suggest improving the system requirements section by actually stating the system requirements. To me the mention of Docker without other details is a black box that I cannot have any intuition for.
- kyrylo 11mo agoI appreciate the feedback. I’ve updated the information to remove the WordPress reference and clarified that the OS must support containerization. Thanks!
- ksec 11mo agoAs much as I like the ONCE idea, even 37signals have found it hard to be sustainable. I think they intend to have an improved version of ONCE idea coming out later. Other than that I hope this project succeed. May be another idea is that you offer paid hosted model as well as the ONCE product. This is similar to a lot of OSS web analytics. People are happy to pay for subscription / services as long as they know they are not locked in. i.e They are happy to pay for $20 - even $2K per month due to Opex rather than Capex reason or saving the manual labour hassle. And they have the choice to move their data to ONCE product when they see fit. All the best I hope this work out.
- kyrylo 11mo agoI follow 37signals updates on ONCE, but I don’t recall them mentioning sustainability issues. Could you share a link? Thanks for the suggestion! A hosted solution could definitely be an option. That said, some people have mentioned that having an indie dev behind it can be a turn-off. I’ll keep exploring ideas that make sense, though.
- jamie_ca 11mo agoI want to vaguely recall that they had like 4 products in planning stages for ONCE, released Campfire, then released Writebook for free, and then made Campfire free. Lately on the podcast I only hear about Fizzy and one other unannounced SAAS that they're putting dev energy behind, nothing about other ONCE products. I suspect that without a critical mass of usage driving word of mouth the long tail of sales was basically nil, and long-term even fairly small products weren't looking to recoup dev costs any time soon.
- kyrylo 11mo agoFizzy uses SQLite for multitenancy. I suppose that’s why it takes up so much of their time. Being a fan of SQLite, I hope this approach succeeds. P.S. I know Campfire has made “six figures”