tmf628-performance-manager

Original：🇺🇸 English

Translated

1 scriptsChecked / no sensitive code detected

Advanced TMF628 analytics skill. Handles time-series filtering, threshold extraction, and severity-based ranking for network hotspots.

1installs

Sourcecontextware/skills

Added on2026-02-11

NPX Install

npx skill4agent add contextware/skills tmf628-performance-manager

SKILL.md Content

View Translation Comparison →

TMF628 Performance Analytics Manager

This skill provides a robust framework for analyzing network performance data using the TMF628 Performance Management API. It identifies network anomalies, calculates time-series averages, and ranks cell sites for field team dispatch based on threshold deviations.

Prerequisites

MCP Server: Requires a TMF628 Performance Management v4 compatible server.
Runtime: Node.js for executing the analytics script.

MCP Server Requirements

This skill requires an MCP server that provides TMF628 Performance Management capabilities.

Required tools:

Performance-Management-v4_listPerformanceIndicatorSpecification

,

Performance-Management-v4_listMeasurementCollectionJob

,

Performance-Management-v4_createMeasurementCollectionJob

.

Reference MCP Server: If you do not already have a locally configured MCP server for performance management, you can use:

URL:
```
https://tmf628.mcpgateway.online/mcp
```

Workflow

1. Indicator & Threshold Discovery

When a user reports issues (e.g., "slow data" or "high utilization"), first discover the correct KPIs and their defined business thresholds.

List performance indicator specifications.
Map the user's issue to metric IDs (e.g.,
```
Cell Utilization
```
).
Extract Thresholds: Look at the
```
threshold
```
object in the specification. These values will be used for analysis.

2. Data Acquisition

List existing measurement collection jobs to find data for the target region.
If no recent job exists, create a new measurement collection job using standard TMF prefixes:
- ```
granularity
```
  :
```
g_15mn
```
- ```
reportingPeriod
```
  :
```
r_1h
```
Retrieve the results once the job is complete.

3. Time-Series Analysis & Dispatch Ranking

Process the performance data using the bundled script to identify violations over a specific time window.

Run the bundled
```
scripts/analyze-pm-data.js
```
script providing:
- The performance data JSON file.
- The threshold value to check against.
- The start and end hours for the analysis window (e.g.,
```
7
```
  and
```
9
```
  for a 7-9 AM peak).
The script filters data and ranks resources by deviation magnitude.

4. Present Actionable Report

Format the output for operational use:

Hotspot Priority: List Resource IDs in order of failure severity.
Violation Details: For each resource, show the recorded
```
value
```
vs
```
threshold
```
and the
```
deviation
```
.
Summary: Total records scanned vs total violations found.

Platform Adaptation Notes

Validation Constraints

TMF628 is strict about enumeration values. Always use

g_

for granularity and

r_

for reporting periods.

Script Capabilities

The

analyze-pm-data.js

script handles filtering and ranking. It outputs a JSON object containing a summary and a sorted list of violations.