vllm-observability SKILL.md skillA
vllm-observability from air-gapped/skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add air-gapped/skills/vllm-observability@git:20261018.5f92473Pins git:20261018.5f92473 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- air-gapped/skills
- Path
.claude/skills/vllm-observability/SKILL.md- Versions recorded
- 11
- Latest version
- git:20261018.5f92473, committed Oct 18, 2026
- First version recorded
- May 29, 2026
- Size
- 12,058 bytes
- Repository stars
- 6
- License
- MIT
Observe production vLLM from the data it already emits — the V1 `/metrics` Prometheus surface, SLO alerts on TTFT/ITL/queue/KV/preemption/aborts/corrupted logits, the shipped Grafana dashboards, OTLP tracing (`--otlp-traces-endpoint`, `--collect-detailed-traces`), DCGM pairing (`SM_OCCUPANCY`, not `GPU_UTIL`), metric renames across upgrades, and triage of slow-TTFT or preemption storms from /metrics alone. Use when building, auditing or fixing vLLM dashboards and alerts, wiring vLLM tracing…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# vLLM observability Target audience: operators running production vLLM on H100/H200 fleets, usually containerized, usually on Kubernetes, on-call for latency and throughput SLOs. ## The core diagnostic rule When something feels slow, read the ratio, not the absolute: | Queue depth | TPOT / ITL | Most likely cause | |---|---|---| | Rising | Stable | **Capacity shortage** — scale out or increase `max-num-seqs` | | Stable | Rising | **Context / model-side** — long-context request, CUDA graph recompile, prefix-cache miss | | Rising | Rising | **Compounding** — usually preemption storm; check `num_preemptions` rate | | Stable | Stable, but TTFT high | **Scheduler stall** — connector (LMCache/NIXL), head-of-line blocking, or engine-core descheduling (ebpf territory) | ## The metric surface in one paragraph …
Read the whole file at its exact version.
How to install
mdr add air-gapped/skills/vllm-observability@sha256:9fc36557b35f7c5fThe line at the top of this page pins the label git:20261018.5f92473, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_sozgyjfxjw6ld4hb)
1 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 11 versions of this file, between May 29, 2026 and Oct 18, 2026, about one every 14 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20261018.5f92473 latest | 2026-10-18 | 5f92473 | 12,058 B | A | view · diff |
| git:20261018.cc96b97 | 2026-10-18 | cc96b97 | 12,212 B | A | view · diff |
| git:20261018.af056ad | 2026-10-18 | af056ad | 12,252 B | A | view · diff |
| git:20261018.5d5da1c | 2026-10-18 | 5d5da1c | 11,555 B | A | view · diff |
| git:20261018.934559a | 2026-10-18 | 934559a | 12,288 B | A | view · diff |
| git:20260922.fd88423 | 2026-09-22 | fd88423 | 14,804 B | A | view · diff |
| git:20260915.0d43610 | 2026-09-15 | 0d43610 | 14,804 B | A | view · diff |
| git:20260915.f93919c | 2026-09-15 | f93919c | 14,801 B | A | view · diff |
| git:20260811.2251198 | 2026-08-11 | 2251198 | 13,736 B | A | view · diff |
| git:20260721.bd7fa37 | 2026-07-21 | bd7fa37 | 13,289 B | A | view · diff |
| git:20260529.06d9c74 | 2026-05-29 | 06d9c74 | 12,062 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (12058 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
air-gapped/skills · 6 stars · license MIT · pushed 2026-11-04 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_sozgyjfxjw6ld4hb GET https://markdownregistry.com/api/v1/resolve?ref=air-gapped/skills/vllm-observability GET https://markdownregistry.com/api/v1/blob/9fc36557b35f7c5f7a9b9e2b4f27417ec6072f807ce7f5d664d9fe59de6472fa
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.