Skip to content
Molt

molt check

Run a collector once, snapshot the result, and report on its health.

Shell
molt check [primary|chaos|c_*]

The building block everything else calls. molt check runs one collector through the real Bright Data CLI, projects the rows, builds a Snapshot, compares it to the baseline, and — if the comparison finds a fault — opens an incident.

What it prints

Shell
molt check chaos
Text
Checking chaos
  collector  c_mt101cvbc0o34ghzh
  target     https://molt-chaos.vercel.app
 
  running…
 
$ bdata scraper run c_mt101cvbc0o34ghzh                    1284 ms
 
  rows  60
  score  ████████░░░░░░░░░░░░  42
  status  BROKEN  2 of 8 fields returned only zeros: comment_count, download_count
 
  field              kind        baseline → now
  comment_count      distorted   60.5 → 0
  download_count     distorted   20,251.5 → 0
 
  incident  i_e956491c-…  detected
  Next: molt watch

If this is the collector's first run, there is nothing to compare against yet — check says so explicitly (BASELINE established) and exits 0 rather than guessing at health from a single sample.

Exit codes

ExitMeaning
0Healthy, or a baseline was just established.
2The collector is broken — a fault was found. CI can gate on this.

Selecting a collector

primary and chaos are the two aliases pinned in this repository's .env (see Onboarding a target for how a third gets added). Any full c_* collector ID also works directly, whether or not it has an alias.

What it does not do

check never proposes a fix. A fault only opens an incident and prints Next: molt watch — the diagnose-and-heal step is a separate, explicit action, so a single check invocation never spends AI-Flow credits by surprise.

PreviousConceptsNextmolt watch