Journal JSON logs

Upload journald exports and other JSON Lines log files with add_journal_json(). Each line becomes one entry in a log channel.

To send logs from a running program instead, see Stream logs.

Connect the client

This guide assumes you have installed the client and setup your API keys.

from nominal.core import NominalClient

client = NominalClient.from_profile("default")

Upload a file

The file needs a .jsonl or .jsonl.gz name and one JSON object per line:

{"__REALTIME_TIMESTAMP": "1789051099000000", "MESSAGE": "Sent ACTUATE_MOTOR command", "level": "INFO"}
{"__REALTIME_TIMESTAMP": "1789051100000000", "MESSAGE": "Motor temperature above target", "level": "WARN"}

• MESSAGE holds the log message.

• __REALTIME_TIMESTAMP holds microseconds since the Unix epoch.

• Every other field, here level, becomes a log argument with a string value.

• Lines without a message or timestamp are skipped.

dataset = client.create_dataset("Engine test logs")
file = dataset.add_journal_json("engine.jsonl", channel="engine_logs")  # channel defaults to "logs"
file.poll_until_ingestion_completed()

Use another timestamp field

Pass both timestamp_column and timestamp_type for a different field or epoch unit:

dataset.add_journal_json(
    "engine-ms.jsonl",
    channel="engine_logs",
    timestamp_column="timestamp_ms",
    timestamp_type="epoch_milliseconds",
)

timestamp_type takes epoch units only. For ISO 8601 or relative times, convert them to epoch values first, or parse the entries and stream them.

View the logs

Add the dataset to an asset or a run. In a workbook on that asset or run, add a Logs panel and select the engine_logs channel.