Log File Anomaly Viewer

Paste a chunk of logs. JSON lines get formatted. ERROR and WARN light up. Stack traces collapse. A filter box hides the noise — all in this tab.

Log anomaly & error highlighter

Pretty-print JSON lines, highlight ERROR/WARN, filter, collapse stack traces. Nothing is uploaded.

Log paste

{"ts":"2026-08-30T13:02:11Z","level":"INFO","msg":"listening","port":8080}
{"ts":"2026-08-30T13:02:12Z","level":"WARN","msg":"retrying payment","attempt":2,"charge_id":"ch_1"}
{"ts":"2026-08-30T13:02:13Z","level":"ERROR","msg":"refund failed","err":"timeout talking to ledger"}
2026-08-30T13:02:13.441Z ERROR refund failed
java.lang.RuntimeException: timeout talking to ledger
	at com.example.RefundService.refund(RefundService.java:88)
	at com.example.RefundService.refund(RefundService.java:41)
	at com.example.ApiHandler.handle(ApiHandler.java:12)
Caused by: java.net.SocketTimeoutException: Read timed out
	at java.base/sun.nio.ch.NioSocketImpl.timedRead(NioSocketImpl.java:283)
	at java.base/sun.nio.ch.NioSocketImpl.implRead(NioSocketImpl.java:305)
{"ts":"2026-08-30T13:02:14Z","level":"DEBUG","msg":"cache miss","key":"refund:ch_1"}
{"ts":"2026-08-30T13:02:15Z","level":"INFO","msg":"request ok","status":200}
Traceback (most recent call last):
  File "worker.py", line 44, in run
    raise TimeoutError("ledger")
TimeoutError: ledger

First 3,000 lines. Heuristic levels and stack folding — not a log shipper. Scrub PII before you share: log anonymizer.

Why a text editor is the slow path

A 4,000-line pod dump in VS Code is searchable. It is not scannable. JSON lines stay minified. A Java stack is 80 frames of beige. The one WARN before the timeout is the line you needed, and it is the same color as everything else.

This viewer is for that paste, not for a production observability stack. It guesses level from JSON fields (level, severity, lvl) or from the text (ERROR, WARN, INFO, DEBUG, FATAL). Stack frames (at com.foo, File ", Traceback, .go:12) attach to the previous event so you can fold them.

Cap is a few thousand lines on purpose. If you need gigabytes, that is jq and a file, not a DOM.

Before you share the paste, run it through the Log Anonymizer. Pretty JSON only is the JSON Formatter.