gsgreen

9 ms·

61.
by gsgreen 4mo ago
We've run nearly 500 benchmark runs: provider selection matters but region does too big time. Still early (3 runs each), so we'll keep adding depth with temporal buffering...
62.
Same Provider. Same $78 Plan. One Hetzner Region Scored 92. Another Scored 14(webbynode.com)
1 points by gsgreen 4mo ago|2 comments
63.
by gsgreen 4mo ago
Hundy %.
64.
by gsgreen 4mo ago
Totally. Availability in US regions seems much tighter than Europe. Ashburn in particular. We'll see if they ramp up (hopefully) sooner than later.
65.
by gsgreen 4mo ago
One thing we've learned doing this: When a result surprises you, the correct response is usually "run it again." Looking forward to running again in later June...
66.
We found a $60 Hetzner VM competing with AWS and Google VMs over $500/mo(webbynode.com)
12 points by gsgreen 4mo ago|5 comments
67.
by gsgreen 4mo ago
Def wasn't a performance story - geography explained it. 2X? Wow.
68.
Why does the exact same Hetzner VM cost $60 in Germany and $119 in Singapore?(webbynode.com)
2 points by gsgreen 4mo ago|1 comments
69.
by gsgreen 4mo ago
Most of the interesting things we've found lately happened because repeated deployments changed and this one def didn't (networking did just a bit). 3 fresh runs of AWS m6i.4xlarge - almost identical. CPU, storage, agentic runtime
70.
$560 AWS m6i.4xlarge VM refused to read the room(webbynode.com)
4 points by gsgreen 4mo ago|1 comments
71.
Our Google Cloud dataset keeps questioning what a $300 VM buys(webbynode.com)
1 points by gsgreen 4mo ago|1 comments
72.
by gsgreen 4mo ago
When first allocating that vm I was utterly shocked.
73.
by gsgreen 4mo ago
Thought this would be a speed story. Ended up being a predictability story. Three fresh deployments looked almost boring. Which is probably exactly what you want from a $512/month VM.
74.
We benchmarked Google Cloud's $512 VM – the speed wasn't the interesting part(webbynode.com)
10 points by gsgreen 4mo ago|6 comments
75.
by gsgreen 5mo ago
What’s weird is the “normal” benchmark layer barely changed - CPU/storage stayed extremely consistent across all 6 runs. The sustained workload behavior was the oddball.
76.
One week later, the same AWS t3.large VM changed personalities again(webbynode.com)
2 points by gsgreen 5mo ago|1 comments
77.
An Asia→Europe Pattern Appeared During Repeated Google Cloud Benchmarks(webbynode.com)
1 points by gsgreen 5mo ago|0 comments
78.
by gsgreen 5mo ago
Not every repeated deployment story ends in weirdness. 6 fresh GCP deployments over 2 weeks. CPU: 702, 704, 703, 702, 704, 703. Storage barely moved. No failures. Boring. In a good way.
79.
We reran this Google Cloud N2 benchmark 6 times over 2 weeks. It barely budged(webbynode.com)
1 points by gsgreen 5mo ago|1 comments
80.
by gsgreen 5mo ago
Run #1 looked boringly dependable. Almost no steal time. Very little drift, etc. Then runs #2 and #3 immediately started behaving differently. Same region. Same plan. 3 fresh deployments. Curious what runs #4–#6 are going to do - will wait
81.
Everyone thinks AWS t3.large is calm. 3 benchmark runs may change things(webbynode.com)
1 points by gsgreen 5mo ago|1 comments
82.
We're seeing distinct runtime lineages emerge inside the same cloud plan/region(webbynode.com)
1 points by gsgreen 5mo ago|0 comments
83.
We kept getting different "personalities" from the same cloud VM SKU(webbynode.com)
1 points by gsgreen 5mo ago|0 comments
84.
by gsgreen 5mo ago
For sure — the interesting part to me isn’t that burstable infra varies. It’s that the variance barely showed up in the normal benchmark layer at all while the sustained workload behavior started separating underneath. Testing for diff scen
85.
Three AWS VPS Runs Looked Identical – One Still Failed Under Load(webbynode.com)
1 points by gsgreen 5mo ago|3 comments
86.
Google Cloud and DigitalOcean Behaved Differently Under Repeated Deployments(webbynode.com)
1 points by gsgreen 5mo ago|0 comments
87.
DigitalOcean NYC looked fine – until we ran it repeatedly (Singapore held up)(webbynode.com)
1 points by gsgreen 5mo ago|0 comments
88.
by gsgreen 5mo ago
Fair if it reads that way but we've been working hard on documenting this data. It's based on repeated runs (3 clusters / 9 runs) on different days — not generated output.
89.
DigitalOcean's NYC region looked fine – until we ran it again(webbynode.com)
3 points by gsgreen 5mo ago|2 comments
90.
by gsgreen 6mo ago
Worth noting: this only shows up with REPEATED runs. Single benchmarks often look completely normal. That’s why a ton of comparisons/benchmarks miss it. Doing this because that drives me a little nuts.

More ›