The log, IDs and names

The log

.itero/itero.log holds everything the daemon does, one JSON object per line:

{"time":"2026-10-21T13:43:24Z","workspace":"01a0f3…","daemon":"01a15f2c…","event":"run.started","workflow":"order","run":"01a15eb3…","run_name":"misty-lake","message":"01a15eb2…"}
{"time":"2026-10-21T13:43:25Z","workspace":"01a0f3…","daemon":"01a15f2c…","event":"step.waiting","workflow":"order","run":"01a15eb3…","run_name":"misty-lake","step":"#check-payment","output":{"payment_ref":"pay-1"}}

Every line has time, workspace, daemon and event. A line about a run has workflow, run and run_name, and "internal": true when its workflow is internal.

`event`	Also
`daemon.started`	every field of the workspace and the daemon
`daemon.stopped`
`message`	`message`, `type`, `data`, and what it did: `started` and `reached` (runs), or `nothing`
`run.started`	`message`, or `schedule`
`step.started`	`step`
`step.done`, `step.waiting`	`step`, `output`
`step.failed`	`step`, `why`
`step.skipped`	`step`, `why` (its `on` did not hold)
`run.note`	`note`: what the run did with a message or a timer
`run.ended`	`status` (`done`, `failed`, `error`, `split`), `why`
`run.skipped`	a schedule came due and the workflow’s `on` did not hold
`source.started`, `source.ended`, `source.gave-up`	`source`, `why`
`paths.watched`	`paths`: every path watched now
`error`	`why`: what went wrong in the daemon itself

It holds what steps printed, never their environment, so no secret. Reading it is jq’s job:

jq 'select(.run_name == "misty-lake")' .itero/itero.log      # one run
jq 'select(.event == "step.failed")' .itero/itero.log        # what broke
jq -r '"\(.time) \(.event) \(.workflow // "") \(.step // "")"' .itero/itero.log
tail -f .itero/itero.log | jq .                             # live

Past its log limit (10 MB) the log is rolled: itero.log becomes itero.log.1, and so on, keeping logs of them (3). See The daemon’s own workflows.

IDs and names

itero gives each of these an identity nobody edits, kept in the journal:

	`id`	`name`	also
the workspace	a UUIDv7, made with its journal	two random words: `quiet-river`	when it was made, its folder, the itero that made it
a daemon’s life	a UUIDv7, at each start	two random words	when it started, its pid, host and itero version
a workflow	a UUIDv7, the first time its file is seen	its file name	its runs so far, its last run
a run	a UUIDv7, at its start	two random words: `misty-lake`	its workflow and status

A name is to say aloud and type; the id stays the identity, and two with one name are told apart by it. A workflow keeps its id while its file name stays, through any edit. Deleting .itero/ starts a new history, with a new workspace id.

Seeing them

itero workflows            # each workflow: what starts it, its runs, the last; the daemon
itero info                 # the workspace: id, name, [meta], limits; and its daemon
itero runs                 # every run: id, name, workflow, status, started, ended
itero runs misty-lake      # one run, by name or id, step by step
itero check                # each workflow: its id, meta, runs so far, last run

itero runs <run> tells a run’s story from the journal: how it started, each step and how it went, the messages it took, where it waits, and its state at the end. Of two runs with one name it shows the newest, and says so.