Real-time vitals. Completely agentless.
Watch your server breathe. Rubuz inspects host kernel statistics, cgroup v2 hierarchies, and container process trees directly over secure SSH. Zero bulky background daemons, zero resource drain, and zero third-party agent vulnerabilities.
| Service / Container | CPU Quota | Memory Resident | Disk I/O | Status | OOM Risk |
|---|---|---|---|---|---|
| api-gateway (Node 22) | 12.4% (0.5 core max) | 312 MB / 1024 MB | 0.2 MB/s | ● Healthy (p99: 14ms) | 0.0% (Safe) |
| postgres-16 (Database) | 4.8% (2 cores max) | 1,240 MB / 2048 MB | 12.1 MB/s | ● Buffer Hit 99.4% | 0.0% (Safe) |
| redis-cache (In-Memory) | 1.2% (0.5 core max) | 128 MB / 512 MB | 0.1 MB/s | ● Hit Rate 98.2% | 0.0% (Safe) |
| worker-queue (Background) | 8.6% (1 core max) | 420 MB / 1024 MB | 2.4 MB/s | ● 4 jobs/sec | 0.0% (Safe) |
Why heavy monitoring agents drag down your VPS. And why agentless wins.
Legacy monitoring tools force you to install invasive background daemons that eat server memory, execute as root, and open external network attack vectors.
The Resource-Draining Agent Model
- ✕300MB - 800MB Resident RAM Tax: Monitoring agents (Datadog, New Relic, heavy collectors) consume significant memory just standing by, starving your actual database and apps.
- ✕Root Daemon Security Vulnerability: Agents often demand root system access and open extra local network ports, creating an unmonitored backdoor and attack surface.
- ✕Telemetry Exfiltrated to Third-Party SaaS: Process lists, internal IPs, and server metadata are continuously uploaded to external proprietary vendor servers.
- ✕High Per-Host Monthly Invoices: You pay $15 to $50/month per host just to see basic CPU and memory charts that the Linux kernel already computes for free.
True Agentless Kernel Telemetry
- ✓0 MB Resident Memory on Host: Zero persistent daemons. Rubuz samples the kernel's native `/proc` and `cgroup v2` tree via existing secure SSH connections only on demand.
- ✓Zero Open Ports & Zero Root Escape Surface: No extra listening ports, no telemetry collector daemon, and no background binaries executing unattended.
- ✓Sub-Second Cgroup V2 Granularity: View exact CPU throttling, memory ceiling limits, and NVMe disk IOPS for individual microservices and database containers.
- ✓100% Included Free Forever: Full live vitals, container metrics, and threshold alert integrations (Telegram, Slack, Email) are built into every Rubuz plan.
Deep kernel visibility without compromise.
Engineered to give you enterprise-grade visibility into your hardware, without consuming the hardware you're trying to monitor.
Pure Cgroups v2 Metric Precision
Unlike legacy panels that scrape inaccurate `top` outputs, Rubuz taps directly into modern Linux cgroup v2 accounting. Track per-container CPU cycles, throttled CFS periods, and accurate resident set size (RSS).
Sub-Second Disk IOPS & Network Profiling
A runaway database query thrashing disk I/O can bring your entire VPS to a crawl. Rubuz tracks read/write IOPS, kernel I/O wait percentages, and ingress/egress packet rates in real time.
Zero Agent Daemon Footprint
No background daemons means your server RAM belongs 100% to your workloads. Whether you run a tiny \$4/mo Hetzner VPS or a 64-core bare-metal machine, Rubuz monitoring leaves zero footprint behind.
Proactive Anomaly & OOM Early Warning
Never wake up to a crashed server. Rubuz analyzes memory growth trajectories and disk capacity trends to warn you hours before an Out-Of-Memory kill happens or disk reaches capacity.
Everything you need to know about Rubuz Telemetry.
How does Rubuz monitor my server without installing a daemon? ↓
Rubuz uses lightweight SSH commands to read directly from the Linux kernel virtual filesystems (`/proc/stat`, `/proc/meminfo`, `/proc/net/dev`) and `/sys/fs/cgroup/`. Because these statistics are already computed continuously by the Linux kernel itself, sampling them takes negligible CPU time (<0.01%) and requires 0 resident memory on your server.
Does real-time telemetry polling slow down my server or apps? ↓
Not at all. In traditional monitoring, Java or Go agent binaries constantly run heavy scraping and serialization loops. In contrast, reading a few lines from the Linux `/proc` filesystem takes less than 2 milliseconds of kernel execution. Your CPU and RAM remain 100% available for your actual customer applications.
Can I see resource metrics for specific containers and databases? ↓
Yes! Rubuz breaks down CPU quota usage, resident memory (RSS), and disk I/O per container. You can immediately identify whether an increase in CPU is caused by your Next.js application, an unindexed PostgreSQL query, or a heavy background worker.
What happens if my server experiences an unexpected memory spike? ↓
Rubuz continuously evaluates memory pressure trends. When memory usage crosses your configured alert threshold (e.g. 85%), you immediately receive an alert via Telegram, Discord, Slack, or Email. Furthermore, rootless container limits ensure that if an individual container runs out of memory, only that specific container is restarted—the rest of your server remains completely online.
Is telemetry data shared with external analytics companies? ↓
Never. Your metrics and operational telemetry remain strictly between your server and your encrypted Rubuz session. We do not sell, monetize, or feed your server data into third-party advertising or analytics networks.
Gain instant visibility. With zero daemon overhead.
Connect any Linux server in 60 seconds. Experience sub-second vitals, per-container metrics, and proactive threshold alerts today.
✓ No credit card required • Connect unlimited servers • 14-day trial