Aarav Dayal

373 posts

Aarav Dayal banner
Aarav Dayal

Aarav Dayal

@aaravdayal

Chief All In Officer @lumatozer_org | 5 years experience in distributed systems

Lumatozer Katılım Ağustos 2024
181 Takip Edilen54 Takipçiler
Sabitlenmiş Tweet
Aarav Dayal
Aarav Dayal@aaravdayal·
Polygon PoS is not even "kind" of decentralized. 1) Firstly to become a validator you need to submit an application. 2) The form link is broken. 3) There can only be 105 validators at a time. 4) 4 miners produced 99.99% blocks in the last 7 days.
English
2
0
0
336
Jake
Jake@JustJake·
Benchmarks make you optimize for the wrong thing Send tweet
English
8
1
39
3.3K
Aarav Dayal
Aarav Dayal@aaravdayal·
@dcolascione what are old method JITs? Also what are tracing JITs? Where can I read books or see lectures about JITs?
English
0
0
0
3
Diptanu Choudhury
Diptanu Choudhury@diptanu·
I am curious about the effects of GC pauses on p95 latencies.
Anurag Goel@anuraggoel

Our @golang load balancer at @render handles more than 150 billion HTTP requests a month across millions of services. The number of times we've wanted to rewrite it in Rust: zero. Go is the most underrated language in infrastructure. "Boring" is the ultimate feature.

English
8
1
132
41.4K
Pablo Salgado
Pablo Salgado@ps1xn·
@mitchellh @morew4rd @thdxr Curious how you dealt with Zig API changes — did you just pin a version or keep upgrading as things changed? Also, how do people using Zig in production usually handle breaking changes? I’m new to Zig, so any tips would be awesome 🙏
English
2
0
2
333
howen
howen@howenyap·
@diptanu dk about p95s but it does cause distributed locks to fail!
howen tweet media
English
5
3
16
2.2K
Aiden
Aiden@WallisDev·
As always, convert it to seconds requestspersecond.fyi
Anurag Goel@anuraggoel

Our @golang load balancer at @render handles more than 150 billion HTTP requests a month across millions of services. The number of times we've wanted to rewrite it in Rust: zero. Go is the most underrated language in infrastructure. "Boring" is the ultimate feature.

English
14
11
725
147.6K
Anurag Goel
Anurag Goel@anuraggoel·
Our @golang load balancer at @render handles more than 150 billion HTTP requests a month across millions of services. The number of times we've wanted to rewrite it in Rust: zero. Go is the most underrated language in infrastructure. "Boring" is the ultimate feature.
English
128
288
4.7K
503.3K
Aarav Dayal
Aarav Dayal@aaravdayal·
@PlanetScale There is an issue with one of your previous articles. Please tell me how to reach out.
English
0
0
0
452
PlanetScale
PlanetScale@PlanetScale·
Sharding is the best way to scale a database. But how does it work? And what's the best way to ensure data and queries are evenly distributed? Our interactive article on database sharding teaches you how.
English
3
25
445
28.7K
Arlan
Arlan@arlanr·
I will send you $100 cash + $500 in @nozomioai credits to whoever solves this. Some of our users can't access trynia.ai. They get SSL_ERROR_RX_RECORD_TOO_LONG. Other domains on the same Vercel project work perfectly. What we know: - Only affects some users, not all - One user's second laptop on the same wifi works, but their main one doesn't even after clearing cache - DNS, SSL cert, Vercel config all valid - Vercel's infra team confirmed it's not on their end Root cause so far: ISPs like Comcast (SafeBrowse) are intercepting TLS handshakes and redirecting to a block page instead of serving our site What we've tried: - Verified DNS, SSL, firewall, bot protection - Switched between A record (76.76.21.21) and CNAME (vercel-dns) - Escalated directly to Vercel engineering Domain is 1 year old, nothing changed on our end, started randomly ~a week ago. Need a fix for EVERY affected user. Not "use a VPN." If you've dealt with ISP-level domain reputation/blocking and know how to get delisted, DM me.
English
27
1
68
19.8K
simeonGriggs
simeonGriggs@simeonGriggs·
@aaravdayal As I understand it, not if there’s overlapping, still-open transactions. If dead tuples are still visible to any open transaction they won’t get vacuumed.
English
2
0
0
76
Aarav Dayal
Aarav Dayal@aaravdayal·
.@simeonGriggs But once task 6 starts stuff deleted during / by task 1 can now be deleted right?
Aarav Dayal tweet media
English
1
0
1
118
itsgalo 🗯
itsgalo 🗯@GaloAndStuff·
terminal aesthetics
Español
45
372
4.4K
94.3K
Aarav Dayal
Aarav Dayal@aaravdayal·
@simeonGriggs Since task 2 and 3 started before task 1 committed/ended. Once they end, then task 1 stuff can be cleaned up. Since task 6 starts after 2 and 3 end, during the execution of task 6, stuff related to task 1 can be cleaned up.
English
0
0
0
25
Aarav Dayal
Aarav Dayal@aaravdayal·
@simeonGriggs I had a slight back and forth with opus 4.6 and here's what I concluded with. I think it will start clearing because any tasks that started before task 1 committed, when they end, then task 1 related stuff can be cleaned up.
English
1
0
0
32
Aarav Dayal
Aarav Dayal@aaravdayal·
@YourAnonOne What if you breathe in really really slowly and breath out very quickly. Is it possible that you could very slightly move? I understand that this is hard cuz of the law of conservation of momentum but I think breathing in very slowly could make a momentum difference overall? idk
English
0
0
1
818
Sushi.com
Sushi.com@SushiSwap·
Sushi Perps are LIVE Powered by @HyperliquidX Early traders get boosted Sushi point multipliers
English
167
139
895
192.3K