All Chapters
Complete table of contents for the melisai documentation. 22 chapters covering Linux performance theory, melisai internals, and production tuning.
Getting Started
| # |
Chapter |
What You'll Learn |
| — |
Quick Start |
Install → run → read results → fix → verify. 2 minutes to first diagnosis |
| 0 |
Introduction |
What melisai is, USE methodology, three-tier architecture, report structure |
| 1 |
Linux Fundamentals |
/proc, /sys, jiffies, CPU states, cgroups v1/v2, PSI, buddy allocator |
| # |
Chapter |
What You'll Learn |
| 2 |
CPU Analysis |
Delta sampling, per-CPU breakdown, load average, CFS scheduler tuning |
| 3 |
Memory Analysis |
MemAvailable vs MemFree, vmstat, PSI, NUMA stats, swap, dirty pages |
| 4 |
Disk I/O Analysis |
/proc/diskstats, 512-byte sectors, I/O schedulers, queue depth |
| 5 |
Network Analysis |
TCP stats, conntrack, softnet, IRQ distribution, NIC hardware, 30+ sysctls |
| 6 |
Process Analysis |
Top-N by CPU/memory, /proc/[pid]/stat, FD counting, state tracking |
| 7 |
Container Analysis |
K8s/Docker detection, cgroup v1/v2, CPU throttling, memory limits |
Internals
| # |
Chapter |
What You'll Learn |
| 8 |
System Collector |
OS detection, uptime, filesystems, block devices, dmesg parsing |
| 9 |
BCC Tools |
67-tool registry, executor, security model, parsers, aggregation |
| 10 |
Native eBPF |
BTF/CO-RE, cilium/ebpf loader, Tier 3 strategy |
| 15 |
Orchestrator |
Two-phase execution, parallel collectors, signal handling, profiles |
| 16 |
Output Formats |
JSON schema, atomic writes, FlameGraph SVG, progress reporter |
Intelligence
| # |
Chapter |
What You'll Learn |
| 11 |
Anomaly Detection |
37 threshold rules, rate-based detection, health score formula |
| 12 |
Recommendations |
35 sysctl fixes, "fix" vs "optimization" types, evidence-based |
| 13 |
AI Integration |
Dynamic prompt generation, 27 anti-patterns, MCP server setup |
Advanced Topics
| # |
Chapter |
What You'll Learn |
| 18 |
GPU & PCIe Topology |
NVIDIA detection, PCI→NUMA mapping, cross-NUMA GPU-NIC pairs, GPUDirect |
| 19 |
Page Reclaim & THP |
Watermarks, direct reclaim, compaction, THP defrag modes, tuning |
| 20 |
NUMA Optimization |
Distance matrix, miss ratio, numactl, K8s topology manager |
Operations
| # |
Chapter |
What You'll Learn |
| 14 |
Report Diffing |
Before/after comparison, USE deltas, histogram changes |
| 21 |
Production Checklist |
All sysctls in one place, one-liner tuning script, anomaly→fix mapping |
Reference
| # |
Chapter |
What You'll Learn |
| 17 |
Appendix |
Glossary, /proc reference, /sys hierarchy, sysctl table, CLI reference |