Skip to content
PayloadPuffin.
Join the waitlist

Your customer says the webhook arrived broken. With recording on, you open the attempt in the console and see the exact request PayloadPuffin sent and the response their server gave. Each destination has two settings that decide whether those two records are kept.

What each attempt records

RecordContains
RequestThe URL, method, headers and body that were sent. The body is the payload after any transform
ResponseThe status, headers, and the first 1,000,000 bytes of the body the destination returned

PayloadPuffin masks credentials in recorded headers. A header named authorization, proxy-authorization, cookie or set-cookie, or one whose name contains token, secret, apikey, api-key or password, keeps only its last four characters, or [redacted] when the value is 12 characters or shorter. webhook-signature is kept whole, because it is what you compare when signature verification fails.

A longer response body is cut at 1,000,000 bytes and marked as truncated. That changes only the record. The attempt's outcome is still the status the destination returned.

Turn recording off or on

Both settings are on for a new destination. Set destination record settings replaces them. A setting you leave out is on:

JSON
{ "storeRequests": true, "storeResponses": false }

It needs the destinations:write scope and answers 200 with the destination. Reset destination record settings turns both back on. In the console, an owner or admin changes the same settings on the destination's page.

A change applies to attempts that start after it. Records already stored stay until their event expires, and turning a setting back on doesn't recover what wasn't recorded in between.

Turn responses off when a receiver echoes data you don't want to keep. Turn requests off when the transformed body holds something you don't want stored a second time. Either way you lose what you'd debug that destination with, so switch it back on when you need to investigate.

These settings cover what PayloadPuffin sends and receives. They don't cover the event you sent: PayloadPuffin keeps that payload regardless, because every delivery is sent from it.

Where you can see the records

The console shows each attempt's request and response, bodies included, on the event's page.

The public API doesn't return record contents. Get a delivery tells you, for each attempt, whether each record exists in records.request and records.response:

stateMeaning
availableThe record was written when the attempt finished. truncated is true if the response body was cut: it was longer than 1,000,000 bytes, or was still arriving when the half-second write window ended
missingThe record was expected but the write failed or there was no room to write it
not_recordedThere will be no record, and reason says why: disabled, not_sent or no_response

not_sent means the attempt failed before a request went out, for example in a transform. no_response means the request went out but no response came back, such as on a timeout.

Recording never changes delivery

PayloadPuffin writes the records of an attempt just before it reports the outcome, and waits for the write for no more than half a second. A storage problem can leave a record missing, but it never fails or repeats a delivery, and a slow write delays the outcome by that half second at most: the write carries on afterwards.

How long records are kept

Records expire with their event. Retention is set per organization and counts from when the event was accepted, not from when the attempt ran, so a retry on the last day is kept no longer than the first attempt. See Retention.