S
snmp161/melisa
Single-binary Linux performance diagnostics via eBPF/BCC, built for humans and AI agents (MCP)
updated 2026-11-19
View on GitHub ↗Direct link · no redirect page
- Star count
- 0 ★ · snmp161/melisa · read from the GitHub search API · (checked 2026-11-19 20:00:48)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-11-20 · last seen 2026-11-20 · seen 1× · listed by 1 sources
- Score
- score 103 · 1 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: Open Source. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
snmp161/melisa;first catalogued 2026-11-20;last seen 2026-11-20。There is no official certification or quality rating here — we record sources and dates only.README excerpt (English, untranslated) · Read the full README on GitHub →
# melisa
Excerpted from the repository README (abridged)
melisa is a point-in-time Linux performance diagnostic tool. It collects data through 8 procfs/sysfs collectors, a registry of 67 BCC tools, and one native eBPF collector, then produces a JSON report with raw measurements, threshold-based anomalies, a heuristic health score, and suggested tuning actions. It works as a standalone CLI or as an MCP server for interactive analysis.
bash
sudo melisa collect --profile quick -o report.json
The command writes a JSON report to report.json (-o - selects stdout). Total runtime depends on the selected profile, host, and available collection tools.
## Key Features
| Feature | What it does |
| Health Score | Heuristic 0–100 summary derived from USE metrics and detected anomalies |
| 37 Anomaly Rules | Static thresholds over current values and, where applicable, two-point rate measurements |
| Recommendations | Suggested tuning commands with evidence, tagged as "fix" or "optimization"; review before applying |
| 67 BCC Tool Specifications | Histogram, event, and stack-trace parsers; actual availability depends on the host |
| GPU/PCIe Topology | NVIDIA GPU metrics, PCI→NUMA mapping, and cross-NUMA GPU-NIC pair detection |
| Page Reclaim Tracking | Direct reclaim rates, compaction stalls, and THP split counters |
| NUMA Analysis | Per-node miss ratio, distance matrix, and CPU mapping |
| Network Data | Conntrack, softnet, IRQ distribution, accept queues, and TCP/UDP sysctls |
| MCP Server | MCP access over stdio for interactive diagnostics |
| Before/After Diff | Comparison of health scores, USE metrics, anomalies, and histograms |
| Observer Effect Mitigation | Tier 1 baselines run before BCC/eBPF tools; PID exclusion and overhead measurement |
| Documentation | 22 chapters in English and Russian covering Linux performance and melisa internals |
## Quick Start
GitHub Releases provides Linux tar.gz archives and DEB/RPM packages for amd64 and arm64. DEB packages target Debian and Ubuntu; RPM packages target AlmaLinux, Rocky Linux, and RHEL. Packages install the binary as /usr/bin/melisa.
bash
# Build (requires Go 1.27+, cross-compile from macOS/Linux)
GOOS=linux GOARCH=amd64 CGO_ENABLED=0 go build -o melisa ./cmd/melisa/
# Deploy and run
scp melisa root@server:/usr/local/bin/
ssh root@server "melisa collect --profile quick -o /tmp/report.json"
# Install BCC tools (first time only)
ssh root@server "melisa install"
The CLI does not require a configuration file. Tier 2 tools are optional and can be installed with melisa install.
## MCP Server
melisa includes a built-in Model Context Protocol server. AI agents connect over stdio and interactively diagnose system performance.
Claude Desktop / Cursor config (claude_desktop_config.json):
json
"mcpServers": {
"melisa": {
"command": "ssh",
"args": ["root@your-server", "/usr/local/bin/melisa", "mcp"]
### MCP Tools
| Tool | What it does |
| get_health | Tier 1 collection with health score and anomalies |
| collect_metrics | Collection using a selected profile, with optional focus and pid filters |
| explain_anomaly | Possible causes and suggested actions for an anomaly ID |
| list_anomalies | All 37 anomaly metric IDs with descriptions |
bash
sudo melisa collect --profile quick -o report.json
The command writes a JSON report to report.json (-o - selects stdout). Total runtime depends on the selected profile, host, and available collection tools.
## Key Features
| Feature | What it does |
| Health Score | Heuristic 0–100 summary derived from USE metrics and detected anomalies |
| 37 Anomaly Rules | Static thresholds over current values and, where applicable, two-point rate measurements |
| Recommendations | Suggested tuning commands with evidence, tagged as "fix" or "optimization"; review before applying |
| 67 BCC Tool Specifications | Histogram, event, and stack-trace parsers; actual availability depends on the host |
| GPU/PCIe Topology | NVIDIA GPU metrics, PCI→NUMA mapping, and cross-NUMA GPU-NIC pair detection |
| Page Reclaim Tracking | Direct reclaim rates, compaction stalls, and THP split counters |
| NUMA Analysis | Per-node miss ratio, distance matrix, and CPU mapping |
| Network Data | Conntrack, softnet, IRQ distribution, accept queues, and TCP/UDP sysctls |
| MCP Server | MCP access over stdio for interactive diagnostics |
| Before/After Diff | Comparison of health scores, USE metrics, anomalies, and histograms |
| Observer Effect Mitigation | Tier 1 baselines run before BCC/eBPF tools; PID exclusion and overhead measurement |
| Documentation | 22 chapters in English and Russian covering Linux performance and melisa internals |
## Quick Start
GitHub Releases provides Linux tar.gz archives and DEB/RPM packages for amd64 and arm64. DEB packages target Debian and Ubuntu; RPM packages target AlmaLinux, Rocky Linux, and RHEL. Packages install the binary as /usr/bin/melisa.
bash
# Build (requires Go 1.27+, cross-compile from macOS/Linux)
GOOS=linux GOARCH=amd64 CGO_ENABLED=0 go build -o melisa ./cmd/melisa/
# Deploy and run
scp melisa root@server:/usr/local/bin/
ssh root@server "melisa collect --profile quick -o /tmp/report.json"
# Install BCC tools (first time only)
ssh root@server "melisa install"
The CLI does not require a configuration file. Tier 2 tools are optional and can be installed with melisa install.
## MCP Server
melisa includes a built-in Model Context Protocol server. AI agents connect over stdio and interactively diagnose system performance.
Claude Desktop / Cursor config (claude_desktop_config.json):
json
"mcpServers": {
"melisa": {
"command": "ssh",
"args": ["root@your-server", "/usr/local/bin/melisa", "mcp"]
### MCP Tools
| Tool | What it does |
| get_health | Tier 1 collection with health score and anomalies |
| collect_metrics | Collection using a selected profile, with optional focus and pid filters |
| explain_anomaly | Possible causes and suggested actions for an anomaly ID |
| list_anomalies | All 37 anomaly metric IDs with descriptions |
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/snmp161-melisa/)
- Source
- github · https://github.com/snmp161/melisa
- Browse by field
- Open source
- added
- today
More entries
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.