Output and exit codes

Console output

Every line starts with the local time. The runner prints:

the command, the project and where each connection points (never secrets);
the log of the run: connecting, reading, progress every 10 seconds on long runs, and the statistics per step;
failed rows: the first 20, each with the row number, the step and the reason (the full list goes to --output);
for preview, every record that would be written;
the summary as the last line, for example Completed with errors: 1,204 row(s) read, 2 failed, 00:01:28.

Problems (errors, warnings and failed rows) go to standard error, everything else to standard output, so 2> errors.txt captures just the problems. --quiet leaves out everything but problems and the summary.

Exit codes

Code	`run` / `preview`	`validate`
0	Succeeded: every row was processed without errors.	No errors (warnings are allowed).
1	Failed or cancelled: the run could not complete, for example a connection failed, a step stopped the run, or Ctrl+C.	—
2	Completed, but some rows failed.	—
3	Nothing was run: invalid command line, project file or connection strings, or the project has errors.	The project or a connection has errors.

Scripts and schedulers use the exit code to tell success from failure:

replisteer run --project accounts.replisteer.json --connections accounts.connections.json --output accounts.log
if %ERRORLEVEL% GEQ 3 echo The project or its connections are wrong & exit /b %ERRORLEVEL%
if %ERRORLEVEL% EQU 2 echo Some rows failed, see accounts.log

Result files

--output <file> saves the result too. The format follows the extension:

Extension	Contents
`.json`	Everything: the summary, exit code, validation issues, status, times, counts, steps, failed rows and preview records. For automation and monitoring.
`.csv`	`run`: the failed rows. `preview`: the records that would be written. `validate`: the problems. Every value quoted, UTF-8 with BOM for Excel.
`.txt`, `.log`	The complete log, as printed (without `--quiet` filtering).

The folder is created when it does not exist. The file is replaced on every run; put a date in the name to keep history (see examples).

JSON result

{
  "runner": "replisteer 1.2026.10.1",
  "command": "run",
  "project": "Accounts nightly",
  "summary": "Completed with errors: 1,204 row(s) read, 2 failed, 00:01:28.",
  "issues": [],
  "exitCode": 2,
  "result": {
    "status": "CompletedWithErrors",
    "startedUtc": "2026-10-20T06:00:02.1234567Z",
    "endedUtc": "2026-10-20T06:01:30.4567890Z",
    "duration": "00:01:28.3333323",
    "wasDryRun": false,
    "rowsRead": 1204,
    "rowsFailed": 2,
    "message": null,
    "failures": [
      {
        "rowNumber": 87,
        "stepId": 1,
        "stepName": "Upsert account",
        "message": "Column 'revenue' expects a number, got 'n/a'",
        "sourcePreview": "AccountNo=C-0087, Name=Northwind, Revenue=n/a"
      }
    ],
    "failuresTruncated": false,
    "steps": [
      { "stepId": 1, "stepName": "Upsert account", "executed": 1204, "skipped": 0, "failed": 2, "recordsAffected": 1202 }
    ],
    "preview": [],
    "durationSeconds": 88.3
  }
}

Field	Meaning
`exitCode`	The exit code of the run.
`issues`	Validation errors and warnings (`severity`, `area`, `message`).
`valid`	`validate` only: TRUE when there are no errors (there is no `result` then).
`result.status`	`Succeeded`, `CompletedWithErrors`, `Failed` or `Cancelled`.
`result.message`	Why a run failed or stopped.
`result.failures`	Failed rows (up to 1,000; `failuresTruncated` is TRUE when there were more).
`result.steps`	Per step: executed, skipped by their condition, failed, records affected.
`result.preview`	`preview` only: the records that would be written.

Last updated on October 29, 2026

Connection strings Windows Task Scheduler