Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
relistan
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
16 ms
·
211.
▲
by
relistan
2y ago
To your point: check out Supermaven. I have no affiliation. But it has way more context than Copilot and is way better at suggestions that make sense in the codebase.
212.
▲
by
relistan
2y ago
You do need it for e.g. database records that are nullable or anywhere else that null/nil is a valid state that differs from zero value. Your alternative approach requires manual work on your part rather than being built in to the bind
213.
▲
by
relistan
2y ago
It’s fairly unpleasant in Go (opinion). Especially for e.g. custom options. YMMV
214.
▲
by
relistan
2y ago
Who could have seen it coming?
215.
▲
by
relistan
2y ago
Would switch permanently
216.
▲
by
relistan
2y ago
Lots of focus in other comments on pedantry. This looks like a cool project.
217.
▲
by
relistan
2y ago
Since 2017, at two different companies, I’ve sent logs via UDP to Sumo Logic, via their collector hosted in our cluster. Sumo Logic is reasonably priced, super powerful, easy to use, and really flexible. Can’t recommend it enough. We do lo
218.
▲
by
relistan
2y ago
This is long overdue. Maybe didn’t go far enough, but a step in the right direction.
219.
▲
by
relistan
2y ago
This. Learning Rust is like a long series of jousting matches with the compiler.
220.
▲
by
relistan
2y ago
Most people off work can’t afford COBRA because the full cost of US health care is so high. COBRA: Nice idea, but useless in practice.
221.
▲
by
relistan
3y ago
Many other places would like to be an outlier like California. Paying more attention to what makes it that way is probably a good place to start.
222.
▲
by
relistan
6y ago
The real gem of the Mesos ecosystem was the much lesser known Singularity scheduler from HubSpot. https://github.com/HubSpot/Singularity I've run this at scale, in production since 2015 and it has been absolutely
223.
▲
The Kernel Change That May Be Slowing Down Your App
(relistan.com)
4 points
by
relistan
7y ago
|
0 comments
224.
▲
Run Envoy in place of docker-proxy and get metrics, tracing for free
(github.com)
1 points
by
relistan
8y ago
|
0 comments
225.
▲
Cgo: When and (Usually) When Not to Use It
(relistan.com)
2 points
by
relistan
9y ago
|
0 comments
226.
▲
by
relistan
10y ago
All punishment is intended as a deterrent, including capital punishment. https://www.law.columbia.edu/law_school/communications/repor... But, even if you don't agree, it's also in no way the cheapest or
227.
▲
by
relistan
10y ago
Minikube doesn't solve the same problems at all. It solves one small part of what Sidecar does. Also note that Sidecar has already been around for 1.5 years and so when you say "rolling your own" it implies that this was a de
228.
▲
Sidecar: Service Discovery for All Docker Environments
(relistan.com)
1 points
by
relistan
10y ago
|
2 comments
229.
▲
by
relistan
13y ago
Or... I could use a language/toolset that makes all that easier. Why does that need further justification?
230.
▲
by
relistan
13y ago
I really appreciate that feedback. Thanks!
231.
▲
by
relistan
13y ago
That wasn't my intention. I didn't decide to like Rust before trying it, I tried it to see if I liked it. I wrote up what I've learned so far. Probably some day this post will seem naive to me, but for now it's the best
232.
▲
by
relistan
13y ago
Thanks for the tip!
233.
▲
by
relistan
13y ago
You're right that it's not an ideal example. I think I said that in the post. The context of the article was the week I spent with Rust and the project I actually did with it. There was not a better example in my code to show. I
234.
▲
by
relistan
13y ago
Disk usage patterns are the problem with most relational stores (and were for Mongo). When you do a query on time series data, you want to look up a start, read some data, and stop when you hit the end point. Getting it laid out largely seq
235.
▲
by
relistan
13y ago
Thanks for the pointers. We are also looking at that as a future plan. We are currently using Thrift which has been rightly pointed out and which I should have mentioned.
236.
▲
by
relistan
13y ago
Good ideas, I'll see if I can generate some time to post more about what we did.
237.
▲
by
relistan
13y ago
An odd oversight. Thanks for pointing it out. I'll see what's up with that.
238.
▲
by
relistan
13y ago
nemothekid is correct. We are not using CQL, we are using Thrift and are running with real columns.
239.
▲
by
relistan
13y ago
It was one of those, "we need to do something now and this will work" solutions. We had a really talented consultant working for us, writing some of the early code. He was familiar with Mongo and wanted to go that route. Early on
240.
▲
by
relistan
13y ago
Hopefully this article wasn't seen as bashing MongoDB. I think I mention it was pretty reliable for us even under load. We were just solving a problem with it for which it wasn't the best solution.
More ›