Support for structured logs?

[johann-petrak]

Basically, log files where each line is a json dictionary with (possibly nested) key-value pairs?

[hamster620]

@johann-petrak Could you provide a sample log file? So that I can create a sample log profile for your reference.

[johann-petrak]

Structured (JSON) logging log files are just jsonl where each line is a JSON dict describing a log event, e.g.

{ "timestamp": "2025-02-17T12:34:56.789Z", "level": "ERROR", "message": "Failed to connect to database", "service": "user-service", "logger": "database.connection", "thread": "main-thread", "host": "server-123", "pid": 45678, "event_id": "abc123xyz", "context": { "user_id": "42", "request_id": "req-5678" }, "stack_trace": "ConnectionTimeoutException: Timeout after 5000ms..." }

The advantage is that arbitrarily nested, structured information can be included which will help with analysis later.
Usually, but not always, certain top level fields in the JSO dict are always present, like "timestamp" or "time", "level" or "severity", or "message", while other fields may vary depending on the logging software, the module getting logged or the log level.

[hamster620]

Structured (JSON) logging log files are just jsonl where each line is a JSON dict describing a log event, e.g.

{ "timestamp": "2025-02-17T12:34:56.789Z", "level": "ERROR", "message": "Failed to connect to database", "service": "user-service", "logger": "database.connection", "thread": "main-thread", "host": "server-123", "pid": 45678, "event_id": "abc123xyz", "context": { "user_id": "42", "request_id": "req-5678" }, "stack_trace": "ConnectionTimeoutException: Timeout after 5000ms..." }

The advantage is that arbitrarily nested, structured information can be included which will help with analysis later. Usually, but not always, certain top level fields in the JSO dict are always present, like "timestamp" or "time", "level" or "severity", or "message", while other fields may vary depending on the logging software, the module getting logged or the log level.

@johann-petrak You can concatenate all unknown JSON properties into one log property such as Extra1, Extra 2 or Summary.

Please refer to attached log profile:
GitHub#55.json

[johann-petrak]

Thank you!

The reason why I was looking at this project is because I was looking for log viewers which

• can display structured logs
• allow to top-down explore nested data within the log fields, i.e. do not just show the JSON data as is but provide a way to navigate the deeply nested information (e.g. trace data)

Cramming the JSON into a single String field would defy both purposes, but I think there are other solutions to my problem.