CLI reference

Every command also accepts -h/--help. This page is generated from the same output, kept in sync by hand against the actual built binary rather than written from memory.

Global flags

Available on every command:

--config string     path to enodia.yaml (default: search standard locations)
--settings string   path to settings.yaml (default: search standard locations; missing is not an error)

enodia collect

collect only gathers facts: an unreachable target is recorded as an
observation with an error, not treated as a command failure. The exit
code reflects whether collection itself ran, not what it found - that
judgement is check's job.

-o, --output string   output file, or - for stdout (default "-")

enodia check

check evaluates the patch, lifecycle and newer-branch axes for every
target as of the moment its data was collected. Without --from it
collects first, in this same process - not a second code path. With
--from it reads an existing inventory and only ever reaches the network
for the lifecycle resolver.

--view selects the table focus: compact (default), lifecycle, drift, or
fleet - the offline-only view of version spread across a product's
instances. When --view is not passed, settings.yaml's render.default_view
applies instead of the compact default, if set.

--fail-on strings   escalate an axis value to a hard failure, e.g. --fail-on=patch:behind (repeatable)
--from string       read an existing inventory instead of collecting
--view string       table view: compact, lifecycle, drift, or fleet (default "compact")
--warn-days int      warn this many days before a lifecycle boundary is reached

enodia export

export writes one self-contained file - nothing here serves it; nginx,
a cron job, or a systemd timer regenerating it is what does.

--format selects json, prometheus, or html. When --format is not passed,
settings.yaml's export.default_format applies instead, if set; the
built-in default stays json either way.

--view restricts an html export to one view (compact, lifecycle, drift,
or fleet) instead of all four stacked sections; ignored by
json/prometheus, which always carry every observation and assessment.

--fail-on strings   escalate an axis value to a hard failure (repeatable)
--format string     output format: json, prometheus, or html (default "json")
--from string       read an existing inventory instead of collecting
-o, --output string  output file, or - for stdout (default "-")
--view string        html only: restrict the report to one view (compact, lifecycle, drift, fleet); default is all four
--warn-days int       warn this many days before a lifecycle boundary is reached

enodia config

Three subcommands, no flags of its own beyond the global ones.

enodia config path

Prints which config file would actually be used, given the current search order — see Configuration.

enodia config resolve

resolve tries https, then http (never http first - that would put a
credential on the wire in the clear), for every target whose address
has no explicit scheme. It sends no credentials and does not modify the
config file - it only reports what scheme each target would use.

enodia config validate

Validates the config file and its credential references — including whether a target’s declared credentials: name actually resolves to an entry the probe understands, entirely offline.

enodia products

Lists every supported product — see Supported products for the full table.

enodia history

history reads every "*.jsonl" file in --dir and evaluates each one
against its own collection time - not today - building one timeline
per target ID. Producing that directory needs no code at all, just
"enodia collect -o \"$(date +%F).jsonl\"" on a schedule. This command is
the other half: reading many of them back as one history.

--dir string         directory of dated *.jsonl inventories (required)
--fail-on strings    escalate an axis value to a hard failure (repeatable)
--format string      output format: table or json (default "table")
-o, --output string   output file, or - for stdout (default "-")
--warn-days int        warn this many days before a lifecycle boundary is reached

enodia serve

serve collects and evaluates on a timer (--interval) and serves
whatever the last successful cycle produced - a request never triggers
a new collection. There is no built-in authentication or TLS: put this
behind a reverse proxy.

Endpoints: / (HTML, all four views), /report.json, /metrics
(Prometheus), /healthz (liveness only - never touches the snapshot).

--fail-on strings     escalate an axis value to a hard failure (repeatable) - reflected in the served report, not a process exit code
--interval duration    how often to refresh the snapshot (default 1h0m0s)
--listen string         address to listen on (default ":8080")
--warn-days int          warn this many days before a lifecycle boundary is reached

enodia version

Prints the enodia version, commit, and build date.

enodia about

Prints the enodia logo plus build information: version, commit, build date, Go version, platform, product (probe) count, license, and repository URL.

enodia completion

Generates a shell autocompletion script. Run enodia completion --help for the specific shell you use (bash, zsh, fish, powershell).