DOCA Telemetry Service (DTS)
Scope note. DTS is an externally productized DOCA service: it is
not one of the DOCA SDK libraries, and its source of
truth is the public DTS guide plus the NGC container / helm-chart pages
and the DOCA Platform (DPF) public docs. This skill covers DTS as
deployed and configured — which manifest or invocation starts it,
which dts_config.ini keys shape it, how to verify it exports. It does
not cover the library side: reading telemetry in your own program is
libs/doca-telemetry (doca-telemetry/SKILL.md), publishing
into DTS over IPC from your own program is
libs/doca-telemetry-exporter (doca-telemetry-exporter/SKILL.md).
Where to start: decide the deployment type before anything
else — it selects the manifest, the config-directory path, the restart
mechanism and the debug commands. Route through
doca-setup ## recognize (doca-setup/TASKS.md#recognize)
(uname -m, lspci -d 15b3:, bf-release) to learn whether you are
on a BlueField Arm image (shape A), a plain host (shape B), or inside a
DPF-managed cluster (shape C), then open
TASKS.md ## configure and follow that shape's
column. State the shape in the answer before the first command, even
when the question implies it ("on my BlueField" is shape A: the
built-in static pod under kubelet, inspected with crictl, not
docker). Every configure, change or debug answer ends with the
verification contract, whose green signal is two /metrics scrapes at
least two update intervals apart showing a counter that advanced
(TASKS.md ## test step 3). If the question is what
can DTS collect and export, and in which mode, start at
CAPABILITIES.md ## Capabilities and modes.
If DOCA / the BlueField OS is not installed yet, route to
doca-setup/SKILL.md first.
Host docker container exits right after start (shape B). Check the
documented host shape and copy the exact invocations from the
version-matched guide's "Host Deployment" section instead of inventing
flags: (1) a -host image tag copied from the NGC doca_telemetry tag
list, never guessed; (2) the one-shot init container
(DTS_CONFIG_DIR=host /usr/bin/telemetry-init.sh) runs and exits 0
before the long-running telemetry-run.sh container; (3) the guide's
host-namespace flags and the config, ipc_sockets, data and log
mounts under /opt/mellanox/doca/services/telemetry/; (4) the documented
run uses --rm, so docker logs finds nothing: read the host-mounted
.../telemetry/log directory or re-run once without -d.
BlueField built-in DTS (shape A) serves nothing on port 9100. Inspect
with crictl pods, crictl ps -a and crictl logs <id> (not docker);
check prometheus= and enable-provider= in
/opt/mellanox/doca/services/telemetry/config/dts_config.ini (edit the file
it links to and allow about 60 s for the respawn). Green signal: two
curl <bluefield-ip>:9100/metrics scrapes at least two update intervals
apart show a counter that advanced. Change one variable per iteration.
DPF-managed DTS (shape C). Change providers and keys in the
DPUService helm values (configMapData), confirm the keys with
helm show values, apply with kubectl apply -f, and check the service
conditions with dpfctl. Never hand-edit the INI on a DPF-managed DPU.
Example questions this skill answers well
The CLASSES of DTS questions this skill is built to answer, each with
one worked example. The class is the load-bearing piece; the worked
example is one instance.
- "Which deployment shape am I in, and what does that change?" —
worked example: "I have a BlueField with DTS already running and a
separate x86 host with no DPU; do I configure them the same way?".
Answered by the shape table in
CAPABILITIES.md ## Pattern overview
- the shape-first step in
TASKS.md ## configure.
- "How do I enable or disable a provider / component?" — worked
example: "turn on
ethtool counters and drop sysfs.eth".
Answered by the provider inventory in
CAPABILITIES.md ## Capabilities and modes
- the
enable-provider= / disable-provider= step in
TASKS.md ## modify (in DPF the same option is
set as key: value under configMapData, not in the INI).
- "Which exporter do I pick and where does the endpoint live?" —
worked example: "Prometheus should scrape the DPU, but the fleet
also pushes to a Fluent Bit forwarder". Answered by the pull-vs-push
exporter table in
CAPABILITIES.md ## Capabilities and modes
- the exporter step in
TASKS.md ## configure.
- "How do I upgrade or pin the DTS version on a given shape?" —
worked example: "the BlueField OS bundles an older DTS; I want the
NGC one". Answered by the tag / chart-version rule in
CAPABILITIES.md ## Version compatibility
- the NGC-yaml download pattern in
TASKS.md ## build, with the generic kubelet
static-pod manifest mechanics in
doca-container-deployment/TASKS.md#configure.
- "How do I attach rack / pod / SU labels to every metric?" —
worked example: "per-rack level labels for a DPF-managed fleet".
Answered by the level-labels model in
CAPABILITIES.md ## Capabilities and modes
- the
level_labels.ini / DPUFlavorTemplate step in
TASKS.md ## modify, with the reprovision
warning from
CAPABILITIES.md ## Safety policy.
- "The container is up but nothing is exported — which layer
failed?" — worked example: "
curl :9100/metrics on the DPU
returns nothing, crictl ps shows the pod running". Answered by
the layer split in
CAPABILITIES.md ## Error taxonomy
- the layered ladder in
TASKS.md ## debug.
- "How do I prove a config change actually took effect?" — worked
example: "I edited
dts_config.ini; did DTS restart and pick it
up?". Answered by the restart-within-60-s contract and the green
signal in CAPABILITIES.md ## Observability
- the bounded loop in
TASKS.md ## test.
When to load this skill
Load this skill for hands-on DTS deployment or configuration work:
- (A) BlueField (built-in static pod): understanding the kubelet static
pod under
/etc/kubelet.d/doca_telemetry_standalone.yaml, its
initContainers (telemetry-init.sh, enable-fluent-forward.sh),
disabling auto-start, or upgrading to a newer NGC image than the
BlueField OS bundles.
- (B) Host (docker): running the NGC
-host image with the
documented namespace / device / volume flag families, the one-shot
init container, and deciding whether --gpus all or
the MST device mapping apply to this host.
- (C) DPF (DPUService / helm chart): authoring a
DPUService for the doca-telemetry
chart, mapping dts_config.ini keys onto configMapData, keeping
the Prometheus port consistent between configMapData and
configPorts, and reading dpfctl describe conditions.
- Configuration (all shapes): sampling interval, providers and
components, exporters (Prometheus pull; Fluent Bit / OpenTelemetry /
Prometheus Remote Write push; storage to
/data), IPC with external
applications, level labels and device labels, the HFT HTTP API.
- Debugging: the container exits at start or restart-loops, the
Prometheus endpoint is empty, a provider shows no counters, the Fluent
Bit destination receives nothing, or a DPF
DPUService never reaches
ApplicationsReady.
Do not load this skill for general DOCA orientation, installing
DOCA or the BlueField OS, installing DPF itself, library-API questions,
or building Grafana dashboards. For those, route via
doca-public-knowledge-map/SKILL.md,
doca-setup/SKILL.md, or the matching libs/
skill.
What this skill provides
This is a thin loader. Substantive material lives in two companion
files:
CAPABILITIES.md — the three deployment shapes and what each fixes
(manifest, config path, restart mechanism, debug tooling); the
inventory-discovery procedure (generated config / rendered ConfigMap
over the example tables) with the provider and exporter catalogues
(pull / push / storage); the
version anchor (image tag / chart version, never the host DOCA
release); the layered error taxonomy (image → manifest → init /
config → providers → exporter → consumer); the observability surface
with its named green signal; the safety policy; and the
public-source pointer.
TASKS.md — step-by-step workflows for the in-scope verbs:
configure, build (assembling the deployment artifact — there is
nothing to compile), modify, run, test, debug, plus a
Command appendix, a Deferred task verbs block, and the
Cross-cutting contracts every change-recommending answer ends with.
The skill assumes the operator has the root / sudo or cluster
credentials the public DTS guide expects, and a target that already
runs a BlueField OS image, a docker-capable host, or a working DPF
cluster. It does not provide any of those prerequisites.
What this skill deliberately does not ship
This skill is agent guidance, not a templates or sample-config
bundle. To keep the boundary clean, it deliberately does not contain —
and pull requests should not add:
- Copy-paste manifests, helm values files, or
dts_config.ini
bodies. The kubelet yaml is published per version on the NGC
container page; the chart's keys come from helm show values; the
INI options come from the live guide for the running version. The
agent quotes the documented flag / key families with their "why"
and points at the exact guide section for the verbatim recipe.
- Container image tags or chart versions as facts. Tags follow the
<dts-version>-doca<doca-version>[-host] shape and are read from the
NGC tag list at answer time; a memorized tag is the canonical
hallucination for a service skill. See
doca-version/SKILL.md.
- Grafana dashboards or Prometheus / Fluent Bit server configs.
The skill names the DTS-side endpoint and the consumer-side target
shape and stops there.
- Credentials, image pull secrets, or cluster credential files.
imagePullSecrets is named as a chart key family; its contents are
the user's.
- A
samples/, templates/, or reference/ subtree of any kind.
A mock or incomplete artifact in this skill's tree, even one labeled
"reference", is misleading: operators will read it as
production-ready.
Loading order
- Read this file first to confirm the question is about DTS as
deployed (not the library API, not DPF installation), and fix the
deployment shape via
doca-setup ## recognize (doca-setup/TASKS.md#recognize).
- For the shape table, providers, exporters, version anchor, error
taxonomy, observability and safety policy, see
CAPABILITIES.md.
- For step-by-step workflows — configure, build, modify, run, test,
debug — see TASKS.md, reading only the column for the
chosen shape.
- Apply the safety overlay in every phase. privileged mode with host
namespaces,
systemctl start dpe, fwctl / mst device exposure, and a
DPF flavor-template change that reprovisions DPUs are the
safety-relevant actions — see
CAPABILITIES.md ## Safety policy.
Related skills
doca-setup/SKILL.md — the front door:
## recognize (doca-setup/TASKS.md#recognize) decides the
deployment shape;
## no-install (doca-setup/TASKS.md#no-install) carries the
tag-selection rule this skill reuses.
doca-container-deployment/SKILL.md
— generic kubelet standalone / static-pod manifest mechanics on
BlueField and the pod-spec do-not-invent rule; the DTS-specific
facts (/etc/kubelet.d/doca_telemetry_standalone.yaml, the NGC
configs/<version>/ yaml download, the crictl pods / crictl ps /
crictl logs inspection) come from the DTS guide and the NGC
container page and live in this skill. Shape A delegates only the
generic mechanics there via
## configure (doca-container-deployment/TASKS.md#configure)
and ## run (doca-container-deployment/TASKS.md#run).
doca-version/SKILL.md — canonical
version-handling rules; DTS overlays "the anchor is the container
tag / chart version, not the host DOCA release".
doca-debug/SKILL.md — the
universal debug-loop contract (doca-debug/CAPABILITIES.md#universal-debug-loop-contract)
and shared
command appendix (doca-debug/TASKS.md#command-appendix) that
TASKS.md ## debug overlays.
doca-hardware-safety/CAPABILITIES.md#safety-policy
— DTS is read-only telemetry, but its deployment touches privileged
containers, dpe, firmware-facing devices and DPF reprovisioning.
libs/doca-telemetry (doca-telemetry/SKILL.md) — the
reader library; "consume telemetry in my program" goes there.
libs/doca-telemetry-exporter (doca-telemetry-exporter/SKILL.md)
— the publisher library; "my app reports but DTS sees nothing" is
split: IPC socket / shared-memory mounts are this skill's, the
doca_telemetry_exporter_* calls are that skill's.
doca-public-knowledge-map/SKILL.md
— URL routing for the public DOCA documentation set;
CAPABILITIES.md ## Public-source pointer
carries the DTS guide, NGC and DPF URLs.
1---2license: Apache-2.03name: doca-dts4description: >5 Deploy and configure the NVIDIA DOCA Telemetry Service (DTS) as it6 is actually run: (A) BlueField (built-in static pod), (B) Host7 (docker) on x86_64 / aarch64, or (C) DPF (DPUService / helm chart)8 — plus9 the shared configuration layer (`dts_config.ini`, providers,10 exporters, level labels) that all three share. The user's11 DEPLOYMENT TYPE decides everything; pick it first. Trigger even12 without "DTS" for "prometheus on the DPU shows nothing on 9100",13 "telemetry container exits right after start", "add ethtool14 counters to the telemetry DPUService", "how do I upgrade the15 built-in telemetry on BlueField", or "level labels per rack/pod".16 Refuse and route elsewhere for installing DOCA itself, the17 `doca_telemetry` / `doca_telemetry_exporter` library API, installing18 the DPF cluster, Grafana / Prometheus server setup or dashboard /19 PromQL authoring, and BFB / BSP imaging — those belong to other20 skills or to external products.21metadata:22 source: https://skillmd.com/skills/nvidia/doca-dts23 kind: service24 doca_version: "3.5"25compatibility: >26 Container-shaped DOCA service. On BlueField Arm it ships inside27 the BlueField OS image (upgradable from the NGC image); on an28 x86_64 or aarch64 host it runs from the NGC `-host` image under29 docker; under DPF it is a `DPUService` rendered from the NGC30 `doca-telemetry` helm chart. The version anchor is the container31 tag / chart version, not the host DOCA install. Never fabricate32 a tag or chart version — read it from the NGC page or33 `helm show values`.34---3536# DOCA Telemetry Service (DTS)3738**Scope note.** DTS is an *externally productized* DOCA service: it is39not one of the DOCA SDK libraries, and its source of40truth is the public DTS guide plus the NGC container / helm-chart pages41and the DOCA Platform (DPF) public docs. This skill covers DTS **as42deployed and configured** — which manifest or invocation starts it,43which `dts_config.ini` keys shape it, how to verify it exports. It does44not cover the library side: reading telemetry in your own program is45`libs/doca-telemetry` (`doca-telemetry/SKILL.md`), publishing46into DTS over IPC from your own program is47`libs/doca-telemetry-exporter` (`doca-telemetry-exporter/SKILL.md`).4849**Where to start:** decide the **deployment type** before anything50else — it selects the manifest, the config-directory path, the restart51mechanism and the debug commands. Route through52`doca-setup ## recognize` (`doca-setup/TASKS.md#recognize`)53(`uname -m`, `lspci -d 15b3:`, `bf-release`) to learn whether you are54on a BlueField Arm image (shape A), a plain host (shape B), or inside a55DPF-managed cluster (shape C), then open56[`TASKS.md ## configure`](TASKS.md#configure) and follow that shape's57column. State the shape in the answer before the first command, even58when the question implies it ("on my BlueField" is shape A: the59built-in static pod under kubelet, inspected with `crictl`, not60`docker`). Every configure, change or debug answer ends with the61verification contract, whose green signal is two `/metrics` scrapes at62least two `update` intervals apart showing a counter that advanced63([`TASKS.md ## test`](TASKS.md#test) step 3). If the question is *what64can DTS collect and export, and in which mode*, start at65[`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes).66If DOCA / the BlueField OS is not installed yet, route to67`doca-setup/SKILL.md` first.6869**Host docker container exits right after start (shape B).** Check the70documented host shape and copy the exact invocations from the71version-matched guide's "Host Deployment" section instead of inventing72flags: (1) a `-host` image tag copied from the NGC `doca_telemetry` tag73list, never guessed; (2) the one-shot init container74(`DTS_CONFIG_DIR=host /usr/bin/telemetry-init.sh`) runs and exits 075before the long-running `telemetry-run.sh` container; (3) the guide's76host-namespace flags and the `config`, `ipc_sockets`, `data` and `log`77mounts under `/opt/mellanox/doca/services/telemetry/`; (4) the documented78run uses `--rm`, so `docker logs` finds nothing: read the host-mounted79`.../telemetry/log` directory or re-run once without `-d`.8081**BlueField built-in DTS (shape A) serves nothing on port 9100.** Inspect82with `crictl pods`, `crictl ps -a` and `crictl logs <id>` (not docker);83check `prometheus=` and `enable-provider=` in84`/opt/mellanox/doca/services/telemetry/config/dts_config.ini` (edit the file85it links to and allow about 60 s for the respawn). Green signal: two86`curl <bluefield-ip>:9100/metrics` scrapes at least two update intervals87apart show a counter that advanced. Change one variable per iteration.8889**DPF-managed DTS (shape C).** Change providers and keys in the90`DPUService` helm values (`configMapData`), confirm the keys with91`helm show values`, apply with `kubectl apply -f`, and check the service92conditions with `dpfctl`. Never hand-edit the INI on a DPF-managed DPU.9394## Example questions this skill answers well9596The CLASSES of DTS questions this skill is built to answer, each with97one worked example. The class is the load-bearing piece; the worked98example is one instance.99100- **"Which deployment shape am I in, and what does that change?"** —101 worked example: *"I have a BlueField with DTS already running and a102 separate x86 host with no DPU; do I configure them the same way?"*.103 Answered by the shape table in104 [`CAPABILITIES.md ## Pattern overview`](CAPABILITIES.md#pattern-overview)105 + the shape-first step in106 [`TASKS.md ## configure`](TASKS.md#configure).107- **"How do I enable or disable a provider / component?"** — worked108 example: *"turn on `ethtool` counters and drop `sysfs.eth`"*.109 Answered by the provider inventory in110 [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)111 + the `enable-provider=` / `disable-provider=` step in112 [`TASKS.md ## modify`](TASKS.md#modify) (in DPF the same option is113 set as `key: value` under `configMapData`, not in the INI).114- **"Which exporter do I pick and where does the endpoint live?"** —115 worked example: *"Prometheus should scrape the DPU, but the fleet116 also pushes to a Fluent Bit forwarder"*. Answered by the pull-vs-push117 exporter table in118 [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)119 + the exporter step in [`TASKS.md ## configure`](TASKS.md#configure).120- **"How do I upgrade or pin the DTS version on a given shape?"** —121 worked example: *"the BlueField OS bundles an older DTS; I want the122 NGC one"*. Answered by the tag / chart-version rule in123 [`CAPABILITIES.md ## Version compatibility`](CAPABILITIES.md#version-compatibility)124 + the NGC-yaml download pattern in125 [`TASKS.md ## build`](TASKS.md#build), with the generic kubelet126 static-pod manifest mechanics in127 `doca-container-deployment/TASKS.md#configure`.128- **"How do I attach rack / pod / SU labels to every metric?"** —129 worked example: *"per-rack level labels for a DPF-managed fleet"*.130 Answered by the level-labels model in131 [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)132 + the `level_labels.ini` / `DPUFlavorTemplate` step in133 [`TASKS.md ## modify`](TASKS.md#modify), with the reprovision134 warning from135 [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy).136- **"The container is up but nothing is exported — which layer137 failed?"** — worked example: *"`curl :9100/metrics` on the DPU138 returns nothing, `crictl ps` shows the pod running"*. Answered by139 the layer split in140 [`CAPABILITIES.md ## Error taxonomy`](CAPABILITIES.md#error-taxonomy)141 + the layered ladder in [`TASKS.md ## debug`](TASKS.md#debug).142- **"How do I prove a config change actually took effect?"** — worked143 example: *"I edited `dts_config.ini`; did DTS restart and pick it144 up?"*. Answered by the restart-within-60-s contract and the green145 signal in [`CAPABILITIES.md ## Observability`](CAPABILITIES.md#observability)146 + the bounded loop in [`TASKS.md ## test`](TASKS.md#test).147148## When to load this skill149150Load this skill for **hands-on DTS deployment or configuration work**:151152- **(A) BlueField (built-in static pod):** understanding the kubelet static153 pod under `/etc/kubelet.d/doca_telemetry_standalone.yaml`, its154 `initContainers` (`telemetry-init.sh`, `enable-fluent-forward.sh`),155 disabling auto-start, or upgrading to a newer NGC image than the156 BlueField OS bundles.157- **(B) Host (docker):** running the NGC `-host` image with the158 documented namespace / device / volume flag families, the one-shot159 init container, and deciding whether `--gpus all` or160 the MST device mapping apply to this host.161- **(C) DPF (DPUService / helm chart):** authoring a `DPUService` for the `doca-telemetry`162 chart, mapping `dts_config.ini` keys onto `configMapData`, keeping163 the Prometheus port consistent between `configMapData` and164 `configPorts`, and reading `dpfctl describe` conditions.165- **Configuration (all shapes):** sampling interval, providers and166 components, exporters (Prometheus pull; Fluent Bit / OpenTelemetry /167 Prometheus Remote Write push; storage to `/data`), IPC with external168 applications, level labels and device labels, the HFT HTTP API.169- **Debugging:** the container exits at start or restart-loops, the170 Prometheus endpoint is empty, a provider shows no counters, the Fluent171 Bit destination receives nothing, or a DPF `DPUService` never reaches172 `ApplicationsReady`.173174Do **not** load this skill for general DOCA orientation, installing175DOCA or the BlueField OS, installing DPF itself, library-API questions,176or building Grafana dashboards. For those, route via177`doca-public-knowledge-map/SKILL.md`,178`doca-setup/SKILL.md`, or the matching `libs/`179skill.180181## What this skill provides182183This is a **thin loader**. Substantive material lives in two companion184files:185186- `CAPABILITIES.md` — the three deployment shapes and what each fixes187 (manifest, config path, restart mechanism, debug tooling); the188 inventory-discovery procedure (generated config / rendered ConfigMap189 over the example tables) with the provider and exporter catalogues190 (pull / push / storage); the191 version anchor (image tag / chart version, never the host DOCA192 release); the layered error taxonomy (image → manifest → init /193 config → providers → exporter → consumer); the observability surface194 with its named green signal; the safety policy; and the195 public-source pointer.196- `TASKS.md` — step-by-step workflows for the in-scope verbs:197 `configure`, `build` (assembling the deployment artifact — there is198 nothing to compile), `modify`, `run`, `test`, `debug`, plus a199 `Command appendix`, a `Deferred task verbs` block, and the200 `Cross-cutting` contracts every change-recommending answer ends with.201202The skill assumes the operator has the root / `sudo` or cluster203credentials the public DTS guide expects, and a target that already204runs a BlueField OS image, a docker-capable host, or a working DPF205cluster. It does not provide any of those prerequisites.206207## What this skill deliberately does not ship208209This skill is **agent guidance**, not a templates or sample-config210bundle. To keep the boundary clean, it deliberately does not contain —211and pull requests should not add:212213- **Copy-paste manifests, helm values files, or `dts_config.ini`214 bodies.** The kubelet yaml is published per version on the NGC215 container page; the chart's keys come from `helm show values`; the216 INI options come from the live guide for the running version. The217 agent quotes the documented flag / key *families* with their "why"218 and points at the exact guide section for the verbatim recipe.219- **Container image tags or chart versions as facts.** Tags follow the220 `<dts-version>-doca<doca-version>[-host]` shape and are read from the221 NGC tag list at answer time; a memorized tag is the canonical222 hallucination for a service skill. See223 `doca-version/SKILL.md`.224- **Grafana dashboards or Prometheus / Fluent Bit server configs.**225 The skill names the DTS-side endpoint and the consumer-side target226 shape and stops there.227- **Credentials, image pull secrets, or cluster credential files.**228 `imagePullSecrets` is named as a chart key family; its contents are229 the user's.230- **A `samples/`, `templates/`, or `reference/` subtree** of any kind.231 A mock or incomplete artifact in this skill's tree, even one labeled232 "reference", is misleading: operators will read it as233 production-ready.234235## Loading order2362371. Read this file first to confirm the question is about DTS *as238 deployed* (not the library API, not DPF installation), and fix the239 deployment shape via240 `doca-setup ## recognize` (`doca-setup/TASKS.md#recognize`).2412. **For the shape table, providers, exporters, version anchor, error242 taxonomy, observability and safety policy, see243 [CAPABILITIES.md](CAPABILITIES.md).**2443. **For step-by-step workflows — configure, build, modify, run, test,245 debug — see [TASKS.md](TASKS.md)**, reading only the column for the246 chosen shape.2474. **Apply the safety overlay in every phase.** privileged mode with host248 namespaces, `systemctl start dpe`, fwctl / mst device exposure, and a249 DPF flavor-template change that reprovisions DPUs are the250 safety-relevant actions — see251 [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy).252253## Related skills254255- `doca-setup/SKILL.md` — the front door:256 `## recognize` (`doca-setup/TASKS.md#recognize`) decides the257 deployment shape;258 `## no-install` (`doca-setup/TASKS.md#no-install`) carries the259 tag-selection rule this skill reuses.260- `doca-container-deployment/SKILL.md`261 — generic kubelet standalone / static-pod manifest mechanics on262 BlueField and the pod-spec do-not-invent rule; the DTS-specific263 facts (`/etc/kubelet.d/doca_telemetry_standalone.yaml`, the NGC264 `configs/<version>/` yaml download, the `crictl pods` / `crictl ps` /265 `crictl logs` inspection) come from the DTS guide and the NGC266 container page and live in this skill. Shape A delegates only the267 generic mechanics there via268 `## configure` (`doca-container-deployment/TASKS.md#configure`)269 and `## run` (`doca-container-deployment/TASKS.md#run`).270- `doca-version/SKILL.md` — canonical271 version-handling rules; DTS overlays "the anchor is the container272 tag / chart version, not the host DOCA release".273- `doca-debug/SKILL.md` — the274 universal debug-loop contract (`doca-debug/CAPABILITIES.md#universal-debug-loop-contract`)275 and shared276 command appendix (`doca-debug/TASKS.md#command-appendix`) that277 `TASKS.md ## debug` overlays.278- `doca-hardware-safety/CAPABILITIES.md#safety-policy`279 — DTS is read-only telemetry, but its deployment touches privileged280 containers, `dpe`, firmware-facing devices and DPF reprovisioning.281- `libs/doca-telemetry` (`doca-telemetry/SKILL.md`) — the282 reader library; "consume telemetry in my program" goes there.283- `libs/doca-telemetry-exporter` (`doca-telemetry-exporter/SKILL.md`)284 — the publisher library; "my app reports but DTS sees nothing" is285 split: IPC socket / shared-memory mounts are this skill's, the286 `doca_telemetry_exporter_*` calls are that skill's.287- `doca-public-knowledge-map/SKILL.md`288 — URL routing for the public DOCA documentation set;289 [`CAPABILITIES.md ## Public-source pointer`](CAPABILITIES.md#public-source-pointer)290 carries the DTS guide, NGC and DPF URLs.