We migrated from newrelic to datadog (for cost reasons LMAO) a while back and I miss NRQL every single day I'm building a dashboard.
I enjoy having everything instrumented and in one spot, it's super powerful, but I am currently advocating for self hosting loki so that we can have debug+ level logs across all environments for a much much lower cost. Datadog is really good at identifying anomalies, but the cost for logs is so high there's a non-trivial amount of savings in sampling and minimizing logging. I HATE that we have told devs "don't log so much" -- that misses the entire point of building out a haystack. And sampling logs at 1%, and only logging warnings+ in prod makes it even harder to identify anomalies in lower environments before a prod release.
last hot take: The UX in kibana in 2016 was better than anything else we have now for rapidly searching through a big haystack, and identifying and correlating issues in logs.
Datadog log ingest isn't too expensive which will still enrich the logs and can dump to S3 (log archive) where you can use Athena/Trino
Log indexing is $$$, though for sure
Curious on Loki cost. When I priced out ELK at a smaller company it didn't come in much cheaper than $0.50/Gi everyone seems to charge (30 day retention, 2 shards, object storage backups). Back when I worked at JPMC, their internal service was also billed right around there.
> What little remains sparking away in the corners of the internet after today will thrash endlessly, confidently claiming “There is no evidence of a global cessation of AI on December 25th, 2025, it’s a work of fiction/satire about the dangers of AI!”;
DUDE I was daydreaming about this the other day, but with postgres. Can I ask why you chose clickhouse? My idea was that every op could be an insert, and then the resulting state would be a transaction and logged? So you'd get basically logged cpu state? Idk cool stuff I'm pumped to dive in tonight.
lol for sure, I'm genuinely curious though, like maybe op is a clickhouse dev? Or maybe they had the cursed idea and decided clickhouse was the best fit for various reasons?
Postgres does not have incremental view maintenance for its materialized views like clickhouse does. This means the implementation strategy would need to be quite different. Perhaps triggers is enough and of course if you use functions then sure, you can make an emulator there too.
Plasma on Ubuntu is the what windows 7 could've been. It's been my daily for a couple years now, with jetbrains tooling and vscode. The only reason I boot back into windows is if I have to work on a .net framework app with visual studio. And Ubuntu is even explicitly supported by dell and Lenovo? It's a no brainer tbh. I'm lucky that my corporate IT is cool with it: I showed them how it supports drive encryption, can join our domain and run our patching software to meet all their 'policy'
I've been pleasantly surprised at how far I can go using Rider to do .NET development on a Mac. I was able to do pretty much everything, including running SQL Server in a container.
Yeah dude! Did you see linqpad is supported on osx now? I used a Mac for a couple years, last gen Intel, when I was doing mostly node work, but I never really got used to it. One of my coworkers though has been doing full-time .net on a Mac with jetbrains for probably about 2 years now and he said he's just as productive as before at this point.
Dude I've been running Ubuntu with Plasma for almost 3 years now as a daily, and it's perfect. It's what windows 7 could've been. Maybe I'm stuck in my ways, but as a dotnet and devops guy, 2020s was the perfect confluence of open source, works on Linux tooling to fully switch over. Rider, datagrip and vscode, and I don't have to deal with docker or wsl anymore. It's beautiful. I only boot into windows now when I have to deal with .net framework OG stuff, and I'm pretty sure I could kill a weekend and get a VM to boot from my windows nvme so I never have to leave.
I think that's how most people work. I watched a colleague use his MacBook for react dev and all of his windows were just...like whatever size and position they opened at, but never full screen? My 3 monitor brain couldn't compute lol.
I have a laptop, 24" centered horizontal centered and a 24" vertical monitor and do a vertical half split for Spotify/teams/shell/outlook, with docs on the laptop screen and ide on the main window full screen. And virtual desktops for design/research, dev and personal.
Sticking with the standard monitor sizes instead of 4k or ultrawide makes screen sharing way simpler as well!
Small gripe, Modern UI design with 10px of padding around everything means most apps and pages HAVE to be full screen to get anything done.
`Logo` and more importantly the various turtle cursor IDEs, probably strikes an emotional chord for a large subset of HN users. It was a common elementary/middle-school age first introduction to programming in the 80s, 90s and early aughts.
I have a pentium II machine in my closet, with a 6gb scsi drive and a zip drive (!), that has a years worth of keyboarding and logo projects saved.
This is a great feature. We've been able to significantly extend the scope and usefulness of our on-prem SQL Cluster for analytics and reporting with PolyBase by building new transactional systems with cheaper postgres, doing ETLs of third-party data to delta tables in azure storage, and then federating access to them with PolyBase so that nobody in the business has to change how they actually query the data. I'm sure in another decade we'll be fully migrated to some cloud platform but for now, federating the queries is a huge win.
Absolutely. I have an 08 accord, an 02 Silverado and a 2019 odyssey for the wife and family. I want the safety and sensing upgrades for the family and road trips, but I get a bench seat and /real/ buttons!
I'm averaging 20-30% of a new car payment keeping them running and replacing big parts, but I would spend double (maybe lol) that so I can keep the experience. If I had a commute longer than a half hour maybe I'd look into a Prius or Leaf? Just waiting for a cts-v wagon to fall into my lap....
Lightbulbs are infuriating now-a-days. I want poe-driven wired IoT bulbs (and downlights, and I suppose zigbee/zwave where poe can't be added in) that dim and color change on the white spectrum (no rgb silliness I will never need to make my lightbulbs green or purple or whatever), with no flicker and high CRI / full spectrum.
And a black box controller that matches output to exterior conditions automatically. No app, just a black box with a wired light sensor.
You can rip my halogen reading lamps cold dead hands.
I enjoy having everything instrumented and in one spot, it's super powerful, but I am currently advocating for self hosting loki so that we can have debug+ level logs across all environments for a much much lower cost. Datadog is really good at identifying anomalies, but the cost for logs is so high there's a non-trivial amount of savings in sampling and minimizing logging. I HATE that we have told devs "don't log so much" -- that misses the entire point of building out a haystack. And sampling logs at 1%, and only logging warnings+ in prod makes it even harder to identify anomalies in lower environments before a prod release.
last hot take: The UX in kibana in 2016 was better than anything else we have now for rapidly searching through a big haystack, and identifying and correlating issues in logs.