Error in user YAML: (<unknown>): found character that cannot start any token while scanning for the next token at line 1 column 1

---

```markdown
# RPMs for Grab Tools
**Author:** Laurence Oberman (<[email protected]>) with assistance from AI (Chat GPT/Gemini) 
**License:** GPLv3

A suite of modern, lightweight telemetry and analysis tools written in Rust, designed for storage, performance, and systems engineers. These tools provide structured sampling of Linux resource usage with built-in dashboard generation.

---

📦 Available RPMs

Package	Version	Platform	Description
procstats_grab	1.0.0-1	el7 / el8	Process-level CPU, Memory, and I/O tracking
serverstats_grab	3.0.0-1	el8	System-wide Disk, CPU, Net, and Mem telemetry

---

🚀 Installation & Setup

Standard RHEL 8 / Linux

Install the RPMs and enable the systemd timers to ensure the gatherers run and restart automatically.

# Install Packages
rpm -Uvh serverstats_grab-3.0.0-1.el8.x86_64.rpm
rpm -Uvh procstats_grab-1.0.0-1.el8.x86_64.rpm

# Enable and Start Server Stats
systemctl enable --now serverstats_grab
systemctl enable --now serverstats_grab_restart.timer

# Enable and Start Process Stats
systemctl enable --now procstats_grab
systemctl enable --now procstats_grab_restart.timer

OpenShift / CoreOS Notes

The serverstats_gather and procstats_gather binaries are built specifically for OpenShift CoreOS.

• No Dependencies: No font dependencies or sidecar containers required.
• Non-Root: Runs as the non-root core user.
• Usage: 1. SSH into the node: ssh core@<node>

2. Start gatherers:

./serverstats_gather -g 10 &
./procstats_gather -g 60 &

Note: Data is written to the current working directory of the core user.

---

🛠 Tool Functionality

1. serverstats_grab

Captures key I/O, CPU, network, and memory stats into a single capture file for browser-based analysis.

• Gather Mode: serverstats_grab -g <interval_seconds>
• Analysis: serverstats_grab -a <capturefile> (Generates SVG/PNG graphs and index.html)
• Playback (Deltas):
• Disk: -pD | CPU: -pC | Memory: -pM | Network: -pN

2. procstats_grab

Scans /proc to record timestamped samples of process and thread activity. It provides insight into resource spikes rather than just cumulative totals.

• Gather Mode: procstats_grab -g 10
• Analysis: procstats_grab -a <csv_file> (Generates top 20 tables and spike charts)
• Playback: procstats_grab -p <csv_file> (Replays CSV as a time-series table)

CSV Output Format

The tool generates a 14-column CSV: ts_epoch, pid, ppid, tid, comm, state, utime, stime, num_threads, vmrss_kb, vm_size_kb, read_bytes, write_bytes, cmdline

---

🧰 Included Helper Tools

Helper Tools

Tool	Purpose

• awr_io_analyze — Intelligent analysis of per-node AWR reports
• blktrace_report.py — Summarizes blktrace per-CPU files
• collectl2diskstats, collectl2diskstats-rhel7 — Convert collectl raw files for new tools
• parse_collectlsZ — Parses collectlsZ files to summarize process data
• convert_nmon.py — Converts nmon output for tabular review
• iostat_report.py — Summarizes complete iostat logs
• sanitize_logs.py — Removes IPs and server names from logs
• short_disk_report.py — Short field view for disk stats playback
• truncate_serverstats.py — Truncates serverstats_grab data files to a focused time window

---

📝 Troubleshooting & FAQ

• Playback Issues: Ensure CSVs were generated by the latest version of procstats_grab to ensure consistent quoting.
• Charts Not Loading: Open the generated index.html in a modern browser to view the dashboard.
• Rust Built: These tools leverage the safety and speed of Rust and the Plotters library for charting.

---

Assistance from both Chatgpt and Gemini Developed by Laurence Oberman ([email protected]). Contributions and PRs are welcome!

---