CLI and export

Your data is in a file you own. That is only worth something if there is a way out of it, and a way in without a browser.

npx monitor stats
npx monitor export --format csv > issues.csv
npx monitor purge --days 30 --yes

Every command reads NUXT_MONITOR_DATABASE_URL and NUXT_MONITOR_STORAGE_DIR when the matching flag is absent, so the same invocation works against SQLite and against an external database.

stats

Counts, the span the data covers, and the most frequent issues.

The dashboard answers this better and with context, which is exactly why this stays a summary rather than growing into a second dashboard. What it is for is the case where the dashboard is not reachable: a cron box, a container with no port published, a server somebody is already logged into.

  /srv/app/.monitor/monitor.db

  issues       124 (37 open)
  events       8134
  span         2026-07-30T09:12:04.881Z … 2026-08-13T16:44:02.117Z

  Most frequent:
       2841  TypeError: Cannot read properties of undefined
        640  Error: Upstream rejected the request

export

Writes to stdout, so it redirects and pipes.

Flag	Meaning
--table	issues (default) or events
--format	json (default) or csv
--days	Only the last N days

# Everything, as JSON
npx monitor export --table events > events.json

# Last week's issues, for a spreadsheet
npx monitor export --days 7 --format csv > week.csv

The dashboard has the same thing behind a download button (GET /api/export) for when there is a browser in the loop.

JSON output parses the columns that hold JSON — context, breadcrumbs, tags — back into structure rather than leaving them as strings. A field that is a string containing JSON forces every consumer to parse twice, and the second parse is the one they forget.

CSV cells are defused

A cell beginning =, +, - or @ is a formula to Excel and Sheets, and error messages are attacker-influenced text. Those cells are prefixed with an apostrophe. This is the export of a security tool; handing somebody a spreadsheet that executes on open would be a poor way to repay them for using it.

purge

Deletes events older than --days, and the issues left with none.

The server applies retentionDays on a timer, so this is for the times that is not enough: a limit lowered after the fact, a database that grew while nobody was looking, a copy taken for support that should not keep last month.

npx monitor purge --days 30        # says what it would delete
npx monitor purge --days 30 --yes  # does it

Without --yes it only counts. This is the one command here that destroys anything, and a flag is cheaper than a restore.

hash-password

npx monitor hash-password

Prints a scrypt hash for monitor.auth.passwordHash, so the plaintext never appears in your config or your build output. Reads from a prompt when no password is given, which keeps it out of shell history.