Expanding a delivery on the event log page now shows each recorded
attempt: attempt number, outcome, status code, duration, error and
response body. Previously a failure rendered as "target: failed" and
diagnosing it meant opening the per-webhook SQLite file by hand.
The response body is cut by SQLite via substr over a blob cast, the
same projection the event body uses, so an oversized stored response
never becomes a Go string. The page reports the cut with a marker.
Response bodies and errors are remote content, so both go through a
new delivery.Redactor that strips the target's own destination URL,
path, query and userinfo, plus the values of credential-shaped
request headers, before rendering. Target configuration keeps
reaching the template only as a TargetView.
A body that reaches the cap is treated as cut whether or not SQLite
is what cut it. The delivery engine stops reading a response at its
own cap, which is the same number of bytes this page renders, and
the row it writes records that cut length as the whole length, so
nothing in the row separates a response that ended at the cap from
one severed there. Such a body goes through RedactCut, which drops
any tail that is a proper prefix of a secret: the remote chooses the
padding in front of a credential it echoes, so it chooses where the
cut falls inside that credential. Its marker says the response
reached the recording limit rather than quoting a total the row does
not know.
Redactors are built from an unscoped target load. Deleting a target
only soft deletes the row while its deliveries survive, and a scoped
load would leave exactly those deliveries rendering unredacted. The
views the page lists stay scoped.
Attempt loading is chunked so the IN clause cannot exceed SQLite's
bound-parameter limit, and a chunk that fails fails the page rather
than rendering the deliveries it covered as never having run. The
page renders at most 20 attempts per delivery, counting what it
leaves out.
static/css/tailwind.css is regenerated with tailwindcss for the
utility classes the new markup uses.
The 8 KB render cap from #135 left storage untouched but no route served
the rest, so a body over the cap was reachable only with filesystem
access to the SQLite files — in a product whose purpose is storing
webhooks so they can be inspected.
GET /source/{sourceID}/logs/{eventID}/body serves the whole body to the
webhook's owner, as application/octet-stream with an attachment
disposition and nosniff. Those are a security control, not formatting:
the bytes come from the public receiver and are handed back inside the
operator's authenticated origin, and the existing CSP would not stop a
stored HTML payload executing there. The truncation marker links to it
only when a body was actually cut.
Accepted deviation, documented rather than glossed: #157's definition of
done asks the route to stream from the row. It buffers whole instead,
because database/sql exposes no incremental handle on a SQLite BLOB and
substr range reads re-materialise the entire column per call — an
earlier revision chunked at 64 KiB and was 11-15x slower for a worse
bound. Three independent reviewers confirmed no streaming path exists.
Independently reviewed three times. Two earlier revisions each asserted
a memory bound the code did not have; the final reviewer measured
2.057x at the ingest cap and pinned the two overlapping allocations from
source — the driver's column buffer and database/sql's convertAssign
clone — confirming the stated "roughly two bodies, and 2x is a floor
not a ceiling" is now accurate, since SQLite's own materialisation sits
outside the Go heap.
The event log rendered stored bodies untruncated. Since buffered
rendering landed (#123) that became resident memory per concurrent
viewer, up to tens of MB, driven by payloads unauthenticated clients
supply to the public receiver.
Bound in the query rather than the template, via
substr(cast(body as blob), 1, ?) plus length(cast(body as blob)), so an
oversized body never becomes a Go string at all. Adds an EventLogView
projection carrying the true byte count, and trims a partial UTF-8 tail
without rewriting bodies that are merely invalid UTF-8.
Independently reviewed. The generated SQL was dumped under GORM DryRun
to confirm the cap is a bound parameter, both casts are present, and no
other path selects the full column; soft-delete scope, ordering and
pagination are unchanged.
Correction to the PR body: its quoted mutation output was produced by
removing the bound from eventLogColumns, not by raising the cap to
1<<30 as the text claimed. The reviewer reproduced the real
mutation and confirmed the tests do catch removal of the bound.
Follow-up #157 restores in-app retrieval of bodies above the cap.
