# ThinkRMS

> ThinkRMS is a unified resource monitoring and alerting platform by ThinkAlike, covering compute, storage, network and GPU accelerator cards with unified collection, visualization, fault prediction and AI optimization. It natively supports NVIDIA and domestic GPUs (Ascend, Hygon, Cambricon) for Xinchuang scenarios.

- Product page: https://thinkalike.com.cn/en/thinkrms.html
- Markdown: https://thinkalike.com.cn/en/thinkrms.md
- Free trial: https://rms.thinkalike.com.cn

## What it is

ThinkRMS gives one platform-wide view of compute, storage, network and accelerator resources — unified collection, visual dashboards, multi-dimensional alerting and AI optimization — delivered offline for government and enterprise intranets.

## Monitoring scope

- **Compute** — per-core CPU, memory, load and process-level metrics with built-in host dashboards and alert templates.
- **Storage & disk arrays** — partition capacity, IO and usage trends; hardware RAID and disk SMART health.
- **Network & switches** — NIC traffic, error packets, concurrent connections and multi-vendor SNMP metrics.
- **GPU accelerators** — utilization, memory, temperature, power and ECC/XID deep metrics for NVIDIA and domestic cards.
- **Fault prediction & self-healing** — tiered alerting, suppression/silencing, 20+ notification channels and configurable auto-run scripts.
- **AI optimization** — LLM-driven root-cause analysis, connecting to domestic LLMs on-premises so data stays in-domain.

## Domestic GPU differentiation

Native collection plus normalization rules converge NVIDIA, Ascend, Hygon, Cambricon, Moore Threads, Iluvatar and MetaX into unified metrics for cross-vendor comparison — dashboards and alerts stay unchanged when card types change.

## Editions

Lite (≤ 50 nodes), Standard (50–500 nodes, domestic GPUs + inspection reports) and Group (multi-datacenter, edge engines + custom big screen). One unified offline installer.

## FAQ

**What is ThinkRMS?**
A unified resource monitoring and alerting platform by ThinkAlike covering compute, storage, network and GPU accelerator cards.

**Can ThinkRMS monitor domestic GPUs?**
Yes. It natively supports NVIDIA plus Ascend, Hygon, Cambricon, Moore Threads, Iluvatar and MetaX, converging them into a unified metric view.

**Do I have to rebuild dashboards when switching GPU vendors?**
No. Dashboards and alerts stay the same — you only adjust normalization rules, with zero business impact.

**Does ThinkRMS support offline intranet deployment?**
Yes. A unified offline installation package enables one-click intranet deployment; data never leaves your domain.