Skip to content

Choose the right command

Start from the question you have. If you do not know where to begin, run wiretrail capture.har auto, which prints the summary and then investigates the most serious findings for you.

I want an overview

Question Command
What is in this capture, and where should I look first? summary
Just investigate it for me auto
Rank the problems, with the evidence for each diagnose
Can this capture answer my question at all? validate
Who talks to which servers? hosts
Which integration or team does a host belong to? subsystems
Which endpoints were called, and how did they answer? endpoints
What happened, in order? timeline

I think something is called too often

Question Command
Which identical calls repeat? duplicates
Which calls were retried after a failure, and how fast? retries
Where is a burst of calls in a short window? storms
Is it paging in a loop, or asking once per item? pagination
Did the app ignore a rate limit? rate-limit
What differs between calls that look the same? diff

I think something failed or was slow

Question Command
Which calls returned errors, and what did the server say? errors
Which routes redirect again and again, or send the client to another host? redirects
Which calls failed first and succeeded on a later try? transitions
Which calls were slowest, and what were they waiting for? slowest
What does the start of the session look like? startup
What was the first failure, and what failed after it? cascade
Did sign-in, tokens or a token refresh go wrong? auth
Is a token expired, or issued for the wrong place? jwt

I want to look inside, or hand it on

Question Command
Show me one request and response in full show-entry
Which responses mention this text? search
What is the value of this field across responses? extract
Give me the entries as data for a spreadsheet or jq export
Write something I can attach to a ticket report
Replay this call curl
Give the backend team what they need to find these calls handoff

I want to compare or enforce

Question Command
What got worse since the last good capture? compare
Does the capture break my own rules? rules
Are required headers missing, or content types wrong? checks

Two things that apply to every command

  • Narrow it. --filter limits a command to matching requests. validate, show-entry and curl <id> ignore it. See Filter entries.
  • Script it. --json prints the result as a stable document, and for most commands the exit code says whether there were findings. Nine commands always exit 0: see Errors and exit codes. See also JSON output.