$12/Year VPS in Frankfurt + $50 Giveaway – Servitro’s Another LET Post!

ejie0404 · September 8

Received the cpu+ssd freebie. Thanks @Servitro

SlowDD · September 8

I'm struggling not to buy the 1 €/m vps. Such a good fallback/routing machine

batra · September 8

@gbzret4d said:

@Servitro said:

@KeyGenMe said:

@Servitro said:

@Hlamtioui said:
Is ur ips of hetzner and contabo ?

No, we have our own subnets.

Nice budget offers, but what’s the deal with just 1TB of bandwidth? Especially in Interwerk…

1TB in 2026 feels completely out of touch, especially when you're advertising 10Gbps ports. Any chance you could bump the bandwidth pool or offer reasonable add-ons?

We do not offer additional bandwidth at the moment.

1TB is not enough

enough for me! 😜

rider · September 8

Welcome back @Servitro!

Servitro · September 8

@SlowDD said:
I'm struggling not to buy the 1 €/m vps. Such a good fallback/routing machine

Please create a ticket or DM me with a screenshot or video.

Servitro · September 8

@rider said:
Welcome back @Servitro!

@ejie0404 said:
Received the cpu+ssd freebie. Thanks @Servitro

Thankyou

emperor · September 8

Thank you for the offer, in the past i had the 20$/y one, which was good. Cheers

NotFoundException · September 8

@Servitro said:

@SlowDD said:
I'm struggling not to buy the 1 €/m vps. Such a good fallback/routing machine

Please create a ticket or DM me with a screenshot or video.

A video of him thinking about buying the 1€/m VPS? Is that one of your kinks?

Servitro · September 8

@NotFoundException said:

@Servitro said:

@SlowDD said:
I'm struggling not to buy the 1 €/m vps. Such a good fallback/routing machine

Please create a ticket or DM me with a screenshot or video.

A video of him thinking about buying the 1€/m VPS? Is that one of your kinks?

haha

unizhu · September 8

cool

berndy2001 · September 8

@harvindersingh said:
Yabs please.

Is this host any good?

found some in the old thread:
9950X https://lowendtalk.com/discussion/comment/4684110/#Comment_4684110
9950X https://lowendtalk.com/discussion/comment/4690028/#Comment_4690028
7443P https://lowendtalk.com/discussion/comment/4748264/#Comment_4748264
7443P https://lowendtalk.com/discussion/comment/4748550/#Comment_4748550

Easton · September 8

ch1ck · September 8

GLWS

Calypso · September 8

@user0990 said:

@gbzret4d said: 1TB is not enough

Yeah, fully agree — even 10TB or 30TB wouldn’t really be enough these days.

All depends on usecase ofcourse. You can put quite a lot of stuff in 1T...

Servitro · September 9

Bump!

SSG · September 9

great offer, that will definitely helps me to learn few things about VPS, Networking and may be DevOps too

mitnick2 · September 9

good one, i ordered

berndy2001 · September 9

@Servitro how long is this offer valid? I need a vm with in 2-3 weeks.

and the mtr in your looking glass is not working.

luna0116 · September 9

Nene

bchot · September 10

for the win!!

mitnick2 · September 10

opened a ticket, almost 24h , no one reply on the issue

Servitro · September 10

@mitnick2 said:
opened a ticket, almost 24h , no one reply on the issue

Don’t worry, it will be resolved soon.

opck · September 10

Cool!

nick_ · September 10

Nice deals! GLWS!

shanewoog · September 10

I wante

mitnick2 · September 10

I am writing this post not to shame or badmouth the provider, but as constructive feedback and to share data from my recent experience. I understand that on a budget
shared-vCPU plan, resources are pooled and contention can happen. However, performance on my current instance has degraded significantly, and since my support ticket hasn't received a response yet, I hope bringing visibility here might help the team notice and look into the node health.

Fair is fair: the storage performance is actually great, but the CPU and network scheduling appear heavily impacted by an abusive "noisy neighbor".

Here are the objective metrics gathered from the instance:
──────
### 1. Server Specs & Location

• Location: Frankfurt am Main, Germany (AS213495 SERVITRO / GHOSTnet)
• Hypervisor: KVM (QEMU Standard PC i440FX, BIOS SeaBIOS)
• Host CPU: AMD EPYC 7443P 24-Core Processor
• Allocation: 1 vCPU, 4GB RAM, Ubuntu 26.04 LTS
──────
### 2. The Good: Disk I/O (Healthy)

Credit where credit is due: storage is performing well and doesn't seem saturated.

• Direct Sequential Write (oflag=direct, 1M blocks): 447 MB/s (315 MB in 0.70s)
• 4K Sync Latency (oflag=dsync): ~9.47 ms / write (~105 IOPS)
• I/O Wait (%wa): Consistently between 0.00% and 0.02%
──────
### 3. The Issue: Severe CPU Steal Time

I ran a historical check from /proc/stat since boot (~17 hours uptime):

• Total CPU ticks since boot: 6,750,630
• Total Steal ticks (st): 700,082
• Average Historical CPU Steal: 10.37% (over 17 continuous hours)

When running a brief 6-second compute task to test responsiveness under load, %st spiked dramatically, showing that the hypervisor is starving this vCPU:

procs -----------memory---------- ---swap-- -----io---- -system-- -------cpu-------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st gu
 2  0      0 1991024  38232 1323072    0    0     4     0 1772 1858 66  5  0  0 29  0
 2  0      0 1991024  38240 1323100    0    0     0   104 1553 1633 54  7  0  0 39  0
 1  0      0 1990296  38240 1323124    0    0     0     0 1561 1474 59  4  0  0 37  0
 2  0      0 1989280  38240 1323148    0    0     0     0 1318 1338 40  5  0  0 55  0  <-- 55% STEAL
 2  0      0 1987244  38240 1323172    0    0     0     0 1592 1623 59  5  0  0 36  0
 2  0      0 1987244  38240 1323196    0    0     0     0 1681 1837 61  4  0  0 35  0

When asking for compute, the guest only receives 40-60% of the cycles, while 35% to 55% is stolen.
──────
### 4. Network: High Jitter & Packet Loss within Frankfurt

Because the vCPU is frequently stalled by the hypervisor, processing incoming network interrupts seems to be delayed, leading to extreme jitter and packet loss even when pinging local Frankfurt Anycast DNS (1.1.1.1 and 8.8.8.8):

──────
### Conclusion / Note to Provider:

I know this is likely not deliberate on the provider's part, but rather one or two tenants on this node running unregulated heavy workloads (e.g. mining or unconstrained continuous CPU tasks) without fair-share limits in place.

I opened a ticket regarding this, but haven't received an update yet. Posting here politely to:

Ask if other users on this Frankfurt node / AS213495 are experiencing similar CPU contention.
Kindly ask the provider team to check host CPU scheduling or balance instances across nodes.

Happy to provide my ticket/account details via PM to any official rep. Thanks!

Servitro · September 10

@mitnick2 said:
I am writing this post not to shame or badmouth the provider, but as constructive feedback and to share data from my recent experience. I understand that on a budget
shared-vCPU plan, resources are pooled and contention can happen. However, performance on my current instance has degraded significantly, and since my support ticket hasn't received a response yet, I hope bringing visibility here might help the team notice and look into the node health.

Fair is fair: the storage performance is actually great, but the CPU and network scheduling appear heavily impacted by an abusive "noisy neighbor".

Here are the objective metrics gathered from the instance:
──────
### 1. Server Specs & Location

• Location: Frankfurt am Main, Germany (AS213495 SERVITRO / GHOSTnet)
• Hypervisor: KVM (QEMU Standard PC i440FX, BIOS SeaBIOS)
• Host CPU: AMD EPYC 7443P 24-Core Processor
• Allocation: 1 vCPU, 4GB RAM, Ubuntu 26.04 LTS
──────
### 2. The Good: Disk I/O (Healthy)

Credit where credit is due: storage is performing well and doesn't seem saturated.

• Direct Sequential Write (oflag=direct, 1M blocks): 447 MB/s (315 MB in 0.70s)
• 4K Sync Latency (oflag=dsync): ~9.47 ms / write (~105 IOPS)
• I/O Wait (%wa): Consistently between 0.00% and 0.02%
──────
### 3. The Issue: Severe CPU Steal Time

I ran a historical check from /proc/stat since boot (~17 hours uptime):

• Total CPU ticks since boot: 6,750,630
• Total Steal ticks (st): 700,082
• Average Historical CPU Steal: 10.37% (over 17 continuous hours)

When running a brief 6-second compute task to test responsiveness under load, %st spiked dramatically, showing that the hypervisor is starving this vCPU:

procs -----------memory---------- ---swap-- -----io---- -system-- -------cpu-------
r b swpd free buff cache si so bi bo in cs us sy id wa st gu
2 0 0 1991024 38232 1323072 0 0 4 0 1772 1858 66 5 0 0 29 0
2 0 0 1991024 38240 1323100 0 0 0 104 1553 1633 54 7 0 0 39 0
1 0 0 1990296 38240 1323124 0 0 0 0 1561 1474 59 4 0 0 37 0
2 0 0 1989280 38240 1323148 0 0 0 0 1318 1338 40 5 0 0 55 0 <-- 55% STEAL
2 0 0 1987244 38240 1323172 0 0 0 0 1592 1623 59 5 0 0 36 0
2 0 0 1987244 38240 1323196 0 0 0 0 1681 1837 61 4 0 0 35 0

When asking for compute, the guest only receives 40-60% of the cycles, while 35% to 55% is stolen.
──────
### 4. Network: High Jitter & Packet Loss within Frankfurt

Because the vCPU is frequently stalled by the hypervisor, processing incoming network interrupts seems to be delayed, leading to extreme jitter and packet loss even when pinging local Frankfurt Anycast DNS (1.1.1.1 and 8.8.8.8):

• Ping to 1.1.1.1 (20 packets):
• Min: 0.91 ms | Avg: 42.11 ms | Max: 247.91 ms | mdev / jitter: 80.14 ms
• Ping to 8.8.8.8 (20 packets):
• Min: 0.92 ms | Avg: 294.94 ms | Max: 819.85 ms | 20% Packet Loss

──────
### Conclusion / Note to Provider:

I know this is likely not deliberate on the provider's part, but rather one or two tenants on this node running unregulated heavy workloads (e.g. mining or unconstrained continuous CPU tasks) without fair-share limits in place.

I opened a ticket regarding this, but haven't received an update yet. Posting here politely to:

1. Ask if other users on this Frankfurt node / AS213495 are experiencing similar CPU contention.
2. Kindly ask the provider team to check host CPU scheduling or balance instances across nodes.

Happy to provide my ticket/account details via PM to any official rep. Thanks!

Hey, Please share your ticket ID with me via DM, and I’ll look into it.

yellow_cat · September 10

@Servitro Hi! Could you please take a look at ticket #XLD-213778?

I've been waiting since September 9 for the promotional CPU/disk upgrade from this offer, but the VPS is still showing 1 vCPU and 25 GB disk.

I also have two older unanswered tickets:

OMG-110434 — Spamhaus SBL issue

HSJ-893315 — original sales/payment ticket where I also requested the promotional upgrade

At least #XLD-213778 would be great) Thanks!

Servitro · September 10

@yellow_cat said:
@Servitro Hi! Could you please take a look at ticket #XLD-213778?

I've been waiting since September 9 for the promotional CPU/disk upgrade from this offer, but the VPS is still showing 1 vCPU and 25 GB disk.

I also have two older unanswered tickets:

OMG-110434 — Spamhaus SBL issue

HSJ-893315 — original sales/payment ticket where I also requested the promotional upgrade

At least #XLD-213778 would be great) Thanks!

Sure.

mitnick2 · September 10

my tickets solved! Quick update:

The issue has been completely resolved! 
  The CPU steal has returned to normal (sitting around 0–1%), and network latency/jitter has fully stabilized. The server is performing great now.
  A big thank you to the provider's support team for taking action and sorting this out, and thank you everyone here for the feedback and suggestions. 
  Appreciate the good support and quick resolution!

Howdy, Stranger!

Categories

In this Discussion

$12/Year VPS in Frankfurt + $50 Giveaway – Servitro’s Another LET Post!

Comments

OMG-110434 — Spamhaus SBL issue

HSJ-893315 — original sales/payment ticket where I also requested the promotional upgrade

OMG-110434 — Spamhaus SBL issue

HSJ-893315 — original sales/payment ticket where I also requested the promotional upgrade

Howdy, Stranger!

Quick Links

Categories

In this Discussion

$12/Year VPS in Frankfurt + $50 Giveaway – Servitro’s Another LET Post!

Comments

OMG-110434 — Spamhaus SBL issue

HSJ-893315 — original sales/payment ticket where I also requested the promotional upgrade

OMG-110434 — Spamhaus SBL issue

HSJ-893315 — original sales/payment ticket where I also requested the promotional upgrade