thinr product

Pattern: thinr product <action> <productId> [args…] [options]

Product commands operate on every device of a product at once. By default they target active (connected) devices only; offline devices are skipped, since the server would reject calls to them anyway.

Common flags

These appear across most fan-out actions:

Flag	Meaning
-j, --json	Emit one envelope with results[], one entry per device (see JSON envelope).
-g, --group <group>	Restrict to devices in an asset group.
-c, --concurrency <n>	Max parallel operations (default: 10).
-a, --all	Include offline devices.
--fail-fast	Stop dequeueing new devices as soon as one fails.

Discovery and status

list

List every product, with enabled/disabled state and script count. Options: -c, --concurrency <n>.

status <productId>

Snapshot of every device in the product: connection status plus monitoring.

Option	Meaning
-w, --watch [seconds]	Refresh continuously (default interval: 5s). Uses the alternate screen buffer, like htop.
-g, --group <group> · -c, --concurrency <n>	As above.

thinr product status edge-gateways --watch 10

Fan-out operations

exec <productId> <command…>

Run a shell command in parallel on every active device.

Option	Meaning
--timeout <seconds>	Per-device timeout (default: 30).
-g, -c, -a, --fail-fast, -j	As above.

Prints a status table (device, status, detail, duration) followed by per-device output blocks. In JSON mode, emits a summary block plus results[] with stdout, stderr, exit code, timedOut flag, and duration per device.

thinr product exec edge-gateways "systemctl restart app" -c 5 --timeout 60

property <productId> <propertyId>

Read a property from every device. Options: -f, --field <path> dot-path extraction, plus -g, -a, -j.

thinr product property edge-gateways temperature -f data.celsius

resource <productId> <resource>

Call a resource on every active device. Options: -i, --input <key=value> (repeatable), -f, --field <path>, plus -g, -j.

push <productId> <localPath> <remotePath>

Upload a local file to every active device in parallel. Options: -g, -c, -a, --fail-fast, -j.

thinr product push edge-gateways ./config.yaml /etc/app/config.yaml

Product properties (admin)

Properties attached to the product itself, not its devices:

Action	Purpose
property-list <productId>	List product property names.
property-get <productId> <propertyId>	Read one.
property-set <productId> <propertyId> <value>	Create or overwrite. The value is parsed as JSON, falling back to a raw string.
property-delete <productId> <propertyId>	Delete (idempotent).

Dashboard metrics

Metrics drive the fleet dashboard for a product:

Action	Purpose
metric-list <productId>	List configured metrics.
metric-set <productId> <name>	Add or update a metric.
metric-delete <productId> <name>	Remove a metric.

metric-set options:

Option	Meaning
-r, --resource <resource>	Required. Product API resource the dashboard invokes.
-l, --label <label>	Human-readable label.
-f, --field <field>	Dot-path to the numeric value inside the response.
-a, --aggregation <agg>	Fleet aggregation: sum, avg, max, min, count, top, none, distribution.
-v, --visualization <viz>	Rendering hint: kpi, bar, sparkline, list.
-i, --interval <seconds>	Dashboard refresh interval.
--unit <unit>	Display unit suffix (e.g. %, devices).

Logs

Named log sources that thinr device logs --source can stream:

Action	Purpose
logs list <productId>	List configured log sources.
logs show <productId> <name>	Show a single source with its full command, untruncated.
logs add <productId> <name>	Add or replace a source.
logs remove <productId> <name> (alias rm)	Remove a source.
logs set-default <productId> <name>	Set the default source.
logs presets	List available rendering presets.

logs add options:

Option	Meaning
-c, --command <cmd>	Required. Shell command the agent executes for this source.
--default	Mark as the active default.
--pattern <regex>	Custom regex with named groups (time, level, msg) for structured rendering.
--preset <name>	Use a named rendering preset instead.

Playbooks

Playbooks stored on the product. See thinr playbook for authoring ad-hoc playbook files.

Action	Purpose
playbook list <productId>	List registered playbooks.
playbook upload <productId> <name> [file]	Upload a YAML (stdin when omitted or -). Validated against the schema first; --skip-validation bypasses, -d, --description <text> overrides the description.
playbook download <productId> <name>	Print the YAML (or -o, --output <file>).
playbook delete <productId> <name>	Delete (idempotent).
playbook run <productId> <name>	Run against one device or the fleet.

playbook run

Requires exactly one of -d, --device <id> or --fleet.

Shared options: -v, --var <key=value> (repeatable, coerced to the declared type), --vars-file <path>, --dry-run (print the resolved plan), --check (contact devices read-only, report what would change), -y, --yes, -j, --json.

Fleet-mode options:

Option	Meaning
--batch-size <n>	Devices attempted in parallel per batch (default: 5).
--failure-threshold <p>	Abort when the cumulative failure rate reaches P% (default: 10).
-g, --group <id> · --filter <key=value>	Device selection.
--include-offline	Include offline devices.

Fleet runs write a persistent JSON report to playbooks/runs/<timestamp>-<name>-<user>.json in the product's file storage: resolved variables, per-device outcomes, and any abort reason.

thinr product playbook run edge-gateways nginx-config --device edge-gw-17 --check
thinr product playbook run edge-gateways nginx-config --fleet --batch-size 3 -y

Partial failures

Fan-out commands don't stop on individual device errors (unless --fail-fast): each device reports its own outcome, and the run itself completes successfully. When scripting, don't rely on the process exit code to detect per-device failures; check summary.failed and results[].ok in the JSON output instead. The top-level ok reflects the command run, not unanimity.