Licenser

pretty nice extension for pi agent that has some really good prompts, tools and model routing capabilities

We created a fun little tool that lets people explore the Otel data (Traces, Logs, Metrics) they send to axiom via the console in a vi-control inspired terminal ui. As a bonus we bumped the free plan to 3 datasets so everyone can take full advantage of the ax :D

It also serves as a nice showcase what kind of tools can be built with using nothing but the public APIs

Hello friend,

I'm one of the maintainers of tremor, happy to get together and talk about rust event processing if you ever want to :)

Anyone want to learn more, come by the CNCF slack, we're in the #tremor channel :)

I'm happy to answer questions to our design.

If there are any unclear parts, I'm happy to help out.

Hi sorry if the post didn't made that clean. You can sign up here: https://try.fifo.cloud/

Or feel free to shoot dm on twitter (@project_fifo) w/ your e-mail and I will make sure you get an invite :)

I'm happy to answer questions to FiFo.cloud if anyone is curious :)

28 billion data points in 50 gigabytes are not impressive for time series use. That's nearly 2 bytes per data point, many time series databases achieve 1 or less byte per data point.

Outlyr has not build DalmatinerDB they've used it and contributed a bit to it.

Because labels or dimensions are not stored in as a value but as a row identifier in most implementations. That results in having to scan the entire row space and look at every row name and see if it matches the lookup.

Storing labels in a row based system (like SQL) allows querying by value, not column name which takes advantage of all optimizations and indexes making it a lot faster.

That said there is nothing forbidding someone to do both, DalmatinerDB, for example, uses a column-based format for metric values but a row-based format (PostgreSQL) for dimensions.

I helped to create that spreadsheet we tried to be as fair as possible and whenever possible link reproducible, verifiable benchmarks (but then again all benchmarks are lies ;).

That's awesome :) thanks for the great work!

Is there a quick start guide for LASP somewhere on the web?

Thank you, friend! You have no idea how much some encouraging words can mean :)

Agreed, the FreeBSD community has been incredibly welcoming despite differences in approach of some which is super refreshing compared to the "just do it that way we've always done it like that" you encounter in other places.

Please don't quote me on this. I think it's not a License question but they're taking a different approach. Linux binary compatibility has been around for a while now, at least longer than modern LX zones (I don't know about old Solaris lx zones).

There are some more subtle differences like the for BSD Linux emulation is a global setting and 'lx jails' are just jails w/ a Linux userland while branded zones are special kind of zones.

Thanks mate :)

Overall quite good, the system is very stable and the performance is good (probably not the fastest but decent).

The only thing to criticise (to a degree) would be that if the manual isn't followed precisely during the installation you can end up with a busted setup.

For why the choice. It's a solid distributed system, the concepts it uses the same principles as FiFo (masterless setup for high availability). Being written in Erlang means it works flawlessly on SmartOS and FreeBSD plus if you already use erlang gives the advantage to be able to look at the code if needed.

The LeoFS team is very quick to respond, works extremely diligent and takes their work serious (which is a big plus).

Even on the test system which gets brutally shut down (aka plugs pulled) about once or twice a week the installation works flawlessly even after a few month of this torment.

Of cause as always YMMV ;)

Hey I'm Heinz, if you have any questions feel free to ask.

Blind guess, lack of clustering and a history of changing file formats so you loose your data every few month.

Mariano is correct, DalmatinerDB is build and maintained by Project-FiFo. Dataloop (the authors of the blog post) is currently the biggest user (as far as I know at least), using it in their SaaS.

They have however been excellent open source citizens and contributed back improvements, bug reports and suggestions.

DalmatinerDB has been around for a while now. I started working on it during the EUC in 2014, and it was released as part of ProjectFiFo the same year I just really suck at marketing ;).

But I'm kind of curious, are you using Dalmatiner directly? And if so what is your use case?

Actually let me put that paragraph fully in context with the reality.

UDP is no longer used (this is outdated sorry for that), the connection is TCP now as it turned out over all the performance was better.

Dataloss can still occur since DalmatinerDB keeps a cache (which other metric stores might also do). A lot of that can be mitigated by using N=2 (or 3) to store data in multiple nodes that will reduce the chance of dataloss significantly. Keeping caches isn't uncommon however, to ensure full consistency it requires a kind of transaction from that goes from client to server to client, which is 'really' costly and I am convinced not worth it for metrics, a few seconds of lost metrics doesn't warrant the cost of that.

The tests were run with TCP, they do not include non stored metrics.

The situation you describe is more an event then it is an metric, and indeed postgres is much better suited for it, if asked that is what I'd recommend for this kind of issue.

What DalmatinerDB was build of is handling high numbers of metrics like CPU usage, memory usage, etc, everything where for every point in time you have exactly 1 value. In that space it can handle millions of metrics (read millions of inserts) at the same time where a system like postgres would stat to have problems in my experience.

This explains a bit on the take of loosing data: https://vimeo.com/148514080

I feel that time series implies different tradeoffs, this is very specifically build for metrics in the sense of application or hardware metrics which usually have a time time dimension but yes there are no general valid definitions

The storage engine is a separate library, it might be of interest for you: https://github.com/dalmatinerdb/mstore

I find it interesting that the SFConservancy argues 'Oracle has a magic wand to make this all go away' by re-licensing ZFS but completely ignores that just as much they could re-license the kernel under a CDDL compatible license. Given that Linux (in the broader sense) is asking for something it'd sound more reasonable for them to accomodate the other parties demands.

What they do is like saying "Hey I want to buy your house, what you don't want to sell it? How about you make it cheaper then so I can afford it!"