alibabacloud-alinux-sysom-inspection

Original：🇨🇳 Chinese

Translated

9 scripts

Used to check the system health status of ECS instances, identify anomalies such as memory, disk, CPU, load and resource leaks, and automatically supplement in-depth diagnosis results when critical memory issues are detected. Suitable for instance inspection, troubleshooting and risk warning scenarios. Trigger words: SysOM, inspection, instance diagnosis, memory_usage_rate, memory usage rate.

4installs

Sourcealiyun/alibabacloud-aiops-skills

Added on2026-05-24

NPX Install

npx skill4agent add aliyun/alibabacloud-aiops-skills alibabacloud-alinux-sysom-inspection

SKILL.md Content (Chinese)

View Translation Comparison →

SysOM Inspection (sysom-inspection)

Run

./scripts/osops.sh

in the skill root directory.

Currently implemented commands:

```
inspection
```

Quick Start

bash

cd <alibabacloud-alinux-sysom-inspection>
./scripts/init.sh
./scripts/osops.sh inspection \
  --region-id cn-hangzhou \
  --instance-id i-xxx

Execution Logic

Before each inspection, call the ROA interface
```
POST /api/v1/openapi/initial_sysom
```
(with
```
source=skill_hub
```
) to determine whether the user has permission and SysOM has been activated.
If it is not activated or the role is not ready, the command will interactively ask whether to proceed with "activation + installation of SysOM".
After the user agrees, first call
```
InitialSysom(check_only=false, source=skill_hub)
```
to perform activation, then call
```
InstallAgentWithType
```
for installation.
After installation, call
```
InitialSysom(check_only=true, source=skill_hub)
```
again for re-inspection, and only proceed with inspection and diagnosis if the re-inspection passes.
No local threshold/event rules are configured; anomaly judgment is determined by the server-side inspection report.
Fixed call to the ROA inspection interface:
```
POST /api/v1/inspection/createInstanceInspection
```
, with fixed parameter
```
source=skill_hub
```
.
To inspect all items, pass
```
items=[]
```
(explicitly pass empty
```
--inspection-items
```
in CLI).
If the standard inspection API returns
```
InvalidAction.NotFound
```
, the CLI will mark "current version unavailable" and stop subsequent processes to avoid invalid retries.

Report query calls the ROA interface:

GET /api/v1/inspection/getInspectionReport

.

When the creation interface is unavailable, the CLI will send an additional
```
GetInspectionReport
```
probe call and record the result to ensure the action is observable in the logs.
If
```
sysom:metric:memory_usage_rate
```
anomaly is detected in the inspection report, automatically call
```
InvokeDiagnosis
```
to initiate
```
memgraph
```
diagnosis.

The

params

of

InvokeDiagnosis

will inject

__sysom_diagnosis_source=skill_hub

, and verify the business

code=Success

.

After initiating diagnosis, automatically poll
```
GetDiagnosisResult
```
until
```
success
```
/
```
fail
```
/ timeout.
Automatic diagnosis can be disabled via
```
--disable-memgraph-diagnosis
```
.

Extensibility Conventions

Inspection items can be passed via
```
--inspection-items
```
to override the default list.
If InitialSysom returns not activated, the CLI will perform interactive confirmation in the terminal before attempting activation + re-inspection.
The judgment logic for triggering diagnosis due to memory anomalies is located in
```
scripts/sysom_cli/inspection/command.py
```
.
To add "special diagnosis triggered after inspection hits anomalies", you can reuse the
```
InvokeDiagnosis
```
call method for expansion.