Absolutely loved this talk between @aaronfrancis.com and Heikki Linnakangas.
I get drawn in by these kinds of deep dives.
Watching someone with 20+ years in Postgres talk about their craft and the work that goes into a piece of software like Postgres and Neon.
It's truly inspiring.
#postgres
To make developers’ lives easier and ensure a great DX, I’m building ApiGaze SDKs for popular API/web frameworks.
To setup ApiGaze:
1. Install the SDK
2. Add two lines of code
Also, the SDKs will be open-sourced to ensure full transparency about what’s being collected
#opensource #buildinpublic
Haha that's simple, love it!
Thanks for sharing 🙌
Alright! After a few hours of setup I think I got a solid starting point.
Hetzner server running everything for now
GH actions to build docker image and publish to ghcr
Dokploy on the server connected to ghcr, one click deploy
Dare I say...it almost feels like a "real" cloud provider😀
#dokploy
That sounds pretty cool!
Do you mind sharing that bash script? For some inspiration.
And btw if you ever feel like exploring some other options I'm kinda starting to like Dokploy :D
Oh cool! Didn't know about lego, will have to check that out.
Using a simple cron job seems simple (as we like it).
Are you using lego cli and crontab?
Not having to think about it is the ultimate goal! :D
I'm not sure but it's likely to reduce volatility and ensure a "fairer" score for businesses with few reviews.
With very few reviews, a single extreme rating can disproportionately affect the overall score.
And they want businesses to actively collect more reviews... because money...😅
Been there and did some research:
"This means that we automatically include the value of 7 reviews worth 3.5 stars each in all TrustScore calculations. As a business collects more reviews, this becomes a smaller factor in the calculation."
Source:
help.trustpilot.com/s/article/Tr...
Biked to a nearby cafe/bakery to restore by sanity 😂 I'm back online
Reading is for the evening and I just worked out...I'm out of options lol 😂
It's a sunny day...maybe I should just go out play outside 🤔
The power is gone?!
Meaning no internet, no using the desktop pc..
I don't even know what to do then...a little crazy to think about how dependent we are.
Well I'll just wait for the food in the fridge to go bad then 😂
Yeah same, Portainer is great. Just feels more for raw container management than anything else.
My main reason to explore others is actually certificate handling. So far easy with Dokploy. What do you do with certificates? 🙌
Deploying some services to Hetzner.
Previously used Portainer for mangement UI.
Seems like Dokploy and Coolify is the new cool kids in town?
Or is this a case of shiny object syndrome?
#buildinpublic
Totally agree - there is some mentions of it here github.com/fork-dev/Tra...
but it's from 2019, so probably not high priority 😅
Awesome Dan! You won't regret it 😀
I guess I'm a influencer now? haha🤣🤣
Apigaze absolutely isn't an alternative to Honeycomb, Datadog, Azure monitor, Grafana etc
It's an alternative for std out
Inspect error rate, latency, performance, exceptions, data in/out. In a simple manner.
Without using days instrumenting your API and building custom dashboards👈
#buildinpublic
I know...crazy of me to even suggest that 😀
I've never actually used the Ctrl+1... shortcuts.
Ctrl+P and searching (like a cave man) seems natural for some reason.
Now I feel stupid... just think of all the time wasted searching for already open tabs... 😮
If those things are old man stuff then I'm definitely old. See you in the retirement home 😅
Change ZoomIt shortcuts.
Trust me, it's worth it 🤣
You definitely should.
The ultimate sign it's good is that I never felt the urge/need to go look for anything else or cursed it 😂
Same! Switched many times - but since I started using fork.dev I never looked back!
Bonus: it developed and maintained by a couple. Him doing the Mac and her the Windows.
Bonus bonus - it's a one time payment.
I'm building the simplest API monitoring tool.
I'm scratching my own itch here. I've been missing a simple opinionated tool for this.
No complex setup, no steep leaning curve.
Add 1-2 lines of code and get insights.
Think @plausible.io for API monitoring.
#buildinpublic #dev
I know people say you should build in public on X - I've contemplated to do it, but ultimately refused the idea...just dosen't feel right.
I'm definitely going to build in public but it isn't going to be on X.
Anyone doing it on LinkedIn?
#buildinpublic #dev
What's a good way to calc bucket size based on time window
| Window Size (min) | Bucket Size (min) |
| <= 30 | 1 min
| <= 120 | 2 min
| <= 180 | 5 min
....
Any nice approach?
FYI I'll be plotting each bucket in a bar chart
#data
That well quite smooth.
Getting the exact number of request with a status code in a time window seems cool.
This data will be used for plotting.
Filter by env and period available.
#dotnet #buildinpublic #api #telemetry
Getting the first API endpoints in place
The first is traffics core metrics.
Nice to be able to get these for arbitrary periods (1h, 1h15m, 6hour, etc) and for specific env
Considering adding filters for specific endpoints or http verbs.
Next is bucketing.
#dotnet #api #analytics #buildinpublic
Impressive well done! When you say 'already' how long did it make to reach this (major) milestone?
Not saying that you should NEVER write your queries yourself - but in my opinion you need a very good reason to do so.
A there is a lot of good reasons - but don't be the 'we-should-always-write-them-by-hand' just because kind of guy. 😃
#dotnet #efcore
Unpopular (biased) opinion.
EF core should be your default
Seems to be a lot of hate towards EF. But let's be honest - it writes better SQL than most of us.
The work done on the performance of EF Core makes it a no-brainer
Maintained a lot of manual written queries, no thanks in most cases.