heisentest → guides

How to open a log file that is too big for your editor

A 40 MB log opens anywhere. A 900 MB log makes Notepad freeze, makes VS Code warn you about tokenization and then crawl, and is bigger than most "online log viewer" upload limits — and uploading production logs to a random website is usually forbidden anyway. Here are the approaches that actually work, including the ones that don't need anything installed.

The classic answers: command-line tools

They stream the file instead of loading it into memory, so size stops mattering:

less +G app.log          # open at the end, scroll freely
grep -n "ERROR" app.log  # find every error, with line numbers
tail -n 5000 app.log     # just the recent past

On Windows, PowerShell's Get-Content app.log -Tail 5000 does the same job. lnav is the strongest terminal option — it merges and colorizes formats. The trade-off with all of these: you get a pipeline, not a picture. Spotting when things went wrong across a two-hour window from grep output is slow, and correlating two files by timestamp by hand is miserable.

The no-install answer: analyze it in the browser, locally

heisentest takes a different route: drop the file onto the page and it is streamed and indexed inside your browser tab. Nothing is uploaded — the page keeps working with wifi off — so NDA'd production logs stay on your machine.

The whole file becomes a timeline; error spikes are visible at a glance.
Filter millions of lines as you type; filter by level once the format is detected.
Stack traces fold into the error that threw them.
Drop a second file and both merge onto one clock for cause-and-effect reading.

Try it now: open heisentest.com and drop your log — or load the built-in sample incident if you just want to see how it reads a file. Free, no account, Apache-2.0 open source.

When you should still reach for something else

Honest limits: heisentest works on text logs, up to roughly the memory a browser tab can hold (single-digit gigabytes in practice — the file is kept as compact bytes, but it is kept). For logs that live on a remote server, pull the file down first or use lnav over ssh. For a standing fleet of services shipping logs continuously, that is what hosted platforms are for.