* Add Agents snippet to README.
Added important information about LiveKit Agents SDK for building Voice AI, including a code example and links to relevant documentation.
* Add tags for snippet
* ingress: add opt-in support for udp:// URL pull ingress
URL pull ingress previously accepted only http, https and srt source
URLs. Add udp:// as an accepted scheme, gated behind a new
`ingress.enable_udp_url_pull` config option that defaults to false, so
unauthenticated UDP sources are not reachable unless the operator opts
in.
Also add a counterfeiter fake for IngressLauncher and a test covering
the scheme validation matrix.
The ingress handler binds a local socket on the caller supplied address
and port instead of connecting out like the http and srt sources do.
Spell out what that means for operators in both the config field doc
comment and config-sample.yaml: caller controlled local port binding,
unauthenticated and spoofable input, and multicast relaying of traffic
on the handler's local network.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: hold signal messages until the ReconnectResponse goes out
Clients take the ReconnectResponse as the first message on a resumed or
migrated in signal connection, anything ahead of it is dropped. Hold
messages back until it has been written.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix: keep queued participant updates on a path that always flushes
Queue only while the participant is not ready, that queue is always
drained by the join or reconnect response. Log a dropped SDP, it leaves
the negotiation waiting until the state machine recovers it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix: flush queued updates whenever the connection opens
Queue participant updates while the handshake is pending again, and give
the signaller a hook that fires when the connection opens, on an explicit
open and on the handshake window expiring, so the queue always drains.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix: drive the handshake window off a timer and read the gate atomically
The window ran only when something asked whether the handshake was
pending, so a connection with nothing else to send held its queue.
Reading the gate under the lock the flush takes closes the race where an
update queued just after a flush.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix: ignore a handshake timeout from a wait that has ended
Stop cannot cancel a timeout that is already running, so tag each wait
and let a timeout act only on its own. Count opens atomically in the
test, the timer fires on its own goroutine.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
The link target was passed through url.PathEscape, which turns the '?' of
turn:host:3478?transport=udp into %3F. The target of a Link header field
is a URI reference and "?transport=" is part of TURN URI syntax
, not a query string, so a publisher that percent-decodes the
target ends up with an unparsable host.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* telemetry: support roomID change for a participant
A room can get a new id while participants are connected. Key stats
workers as map[roomID]map[participantID] so moving a room is a single
map splice, and add reKeyRoom/RoomIDChanged to do the move.
Stats collected before the change are sealed off with the room they
were collected in so they stay attributed to it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* telemetry: close superseded worker on re-key collision
Only one worker can be keyed at (room, participant). If a re-key lands
on a room that already has a worker for the same participant, keep the
one already filed there and close the superseded one so it drains and
is reaped instead of lingering in the flush list.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* telemetry: hand references to the successor on force close
A ReferenceGuard records that it activated some worker, not which one,
so a superseded worker cannot just drop its references - the survivor
would be left with references it never sees released and would never
close. Hand them over instead.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* report a reason on room-ended telemetry
Room.Close already took a ParticipantCloseReason, but OnClose was func()
with no arguments, so the reason was structurally dropped before it could
reach RoomEnded.
Add types.RoomCloseReason and carry it through Close -> OnClose -> RoomEnded,
where it lands on both the analytics event and the room_finished webhook.
Room.Close now takes the room reason and derives the participant reason from
it, so the two can never disagree. ToParticipantCloseReason maps each reason
to exactly what its call site passed before, so no participant-facing
behaviour changes; a table test pins that mapping.
* fix(test): update webhook test for RoomCloseReason
test/ was outside the packages checked before pushing, so this Close call
site was missed. Also assert the reason reaches the room_finished webhook,
which the unit test cannot cover since it runs with a nil notifier.
* Carry worker kind details into agent job tokens.
WorkerRegistration gains a server-controlled KindDetails field that is
passed to BuildAgentToken when assigning jobs, so kind details flow into
the participant join token for the whole session.
Co-authored-by: Cursor <cursoragent@cursor.com>
Signed-off-by: shishir gowda <shishir@livekit.io>
Co-authored-by: Cursor <cursoragent@cursor.com>
* utils: make Median generic, overflow-safe, and add comprehensive
* tests: fix staticcheck unused variable warning in changenotifier_test.go
* tests: switch from assert to require for consistency with existing tests
* trigger ci rerun
In migration cases, dummy receiver trackInfo is used by relay tracks to
set up the receivers and those need the proper track info.
Also check for proper receiver when adding a migrated track.
TestUnsubscribe checked that the changed-notifier observer was gone as
soon as the unsubscribe had settled, but setDesired leaves the
RemoveObserver call to a goroutine of its own and nothing the test waits
on orders against it, so CI caught the assertion running first.
TestSubscribe has the same defect on the unsubscribed callback, which
unmarkSubscribedTo delivers with a bare go while the subscribed one is
called inline.
Wait for both, each with a message of its own, so that a leak that is
real still says which of them broke.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Let the docker-backed service tests be skipped with a flag.
TestMain called log.Fatalf when it could not reach a docker daemon, so
the whole package refused to run without one, including every test in it
that needs no container at all.
Record why docker is unavailable instead, and gate the tests that want a
container on it. A run asks to go without them with -docker=false;
otherwise a missing daemon still fails the package, so an unreachable
daemon stays a broken build rather than a run that quietly covers less
than the last one did.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
pion no longer starts the repair stream reader when a custom BufferFactory
is set, so the mid/rid/rsid extensions were never observed and simulcast RTX
streams were never paired with their primary streams.
Extract the extensions on the buffer write path instead. Migrated publishers
send no extensions at all, so pair those from SimTracks.
Adds an integration test covering both paths, and moves the vnet setup it
shares with the downtrack test into pkg/testutils/vnettest.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
VP9 SVC sends each spatial layer as its own encoded frame,
so a picture carries one trailer per layer, but only the
top layer's last packet has the RTP marker bit set.
Fix https://github.com/livekit/egress/issues/1347
* config: clarify AdvertiseInternalIP applies to explicit node_ip too
The advertise_internal_ip comment in config-sample.yaml said it works
only when use_external_ip is true. That was accurate until
livekit/mediatransportutil#94, which extended AdvertiseInternalIP to
also be respected on the explicit node_ip path (use_external_ip: false
with rtc.node_ip set) — previously that branch always replaced the
host candidate with node_ip, dropping the original local candidate,
regardless of AdvertiseInternalIP.
Update the comment so it reflects both cases: advertise_internal_ip
keeps the original local candidate alongside the mapped one whether
the mapped IP comes from external-IP discovery or from an explicitly
configured node_ip.
* Update config-sample.yaml
First-party livekit modules are co-released and depend on each other,
so bump them together in a single PR instead of one PR per module.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* Reduce locking in media track + telemetry listener on move participant.
Telemetry listener was not getting it from new room on room move.
* update comment
* fix signal bytes reporting
* room aware telemetry listener in media track
* data track telemetry listener
SendData API messages do not have sender ID or sequence number. So, they
were not cached and hence excluded from the reliable caching feature
which is meant to provide reliability of data channel messages between
the time other participants see participant as ACTIVE (which happens on
ICE connected) and data channel being open (DTLS done + data channels
opened).
Add a small cache for that and flush those messages on data channel
establishment.
* Check for ICE connection before closing participant on signal close.
Only close the participant if ICE has not connected by the time signal
source is closed. If ICE had connected, candidates have been exchanged
and link was established. So, it should be resumable.
Waiting for DTLS closed the participant sometimes in the windowa after
ICE connection, but before DTLS finishes and that unnecessarily closed
the participant forcing a full reconnect.
* variable name
* Flush pending signal responses before closing the web socket.
When the request direction of a signalling connection goes away, the
web socket was closed right away. Responses that the participant had
already sent were dropped.
This loses the leave request on migration. The media node writes
leave(RESUME) and closes the signalling connection just after. The
close won the race, so the client saw a plain web socket close with
code 1000 and never got the leave. It then did a full reconnect
instead of a resume.
Now the response pump is signalled first and drains what is pending,
then the web socket is closed. The producer closes the response
source after its last write, so draining till the source is closed is
a complete flush. A deadline bounds the case where the source stays
open.
The web socket is still closed on all paths, so the ping worker does
not leak.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* Give the response pump a margin over the drain deadline.
Both waits used the same timeout and started at about the same time.
So when the drain ran to its deadline, the outer wait could give up at
the same moment and close the web socket while the pump was still
writing. That write failed and the message was lost. It also logged a
timeout even though nothing was stuck.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
- Renovate
config:recommended (config:base is deprecated) and matchPackageNames globs instead of the deprecated matchPackagePrefixes.
Vulnerability alerts get a fast path: 2-day quarantine, no concurrency/hourly/schedule limits.
Go modules are no longer grouped into one "go deps" PR — each gets its own, so a bad bump can be reverted alone. The pion modules stay grouped as a documented exception: they're co-released and interdependent, so individual PRs wouldn't build.
First-party github.com/livekit/** skips the 2-week quarantine.
go.mod's go directive is no longer an update target — the build toolchain is pinned in the Dockerfile instead.
Dockerfile deps get pinDigests; the golang image is ungrouped with separateMinorPatch so a patch and a minor bump are each separately approvable.
Custom manager to bump the builder image's -alpineA.B suffix together with its digest, which the stock docker manager holds fixed.
- Pinning
Both Dockerfiles pin golang and alpine by digest alongside the readable tag.
GOTOOLCHAIN=local so a go.mod bump fails loudly instead of silently downloading a different toolchain.
apk upgrade in the runtime stage — a digest pin plus the 2-week quarantine would otherwise ship base-package CVEs Alpine has already fixed. This relies on a cold layer cache, which holds today because the release workflow configures no buildx cache; there's a comment saying so.
Workflows resolve the Go version from the Dockerfile via .github/scripts/go-version.sh, so tests, releases and images share one toolchain.
- Tools
All four code generators now come from the module graph, and tools/tools.go (the pre-Go-1.24 blank-import idiom) is replaced by go.mod tool directives:
tool how why
goimports go tool lives in x/tools — its own module is the one being selected
gotestfmt go tool zero dependencies, nothing to skew
wire go run pins x/tools v0.24.1; building it in our graph changes its output
counterfeiter go run unchanged, matches its //go:generate directives
The wire distinction is load-bearing. Building wire inside our module raises it from the x/tools v0.24.1 it pins to our v0.48.0, and that module version difference changes what it generates: it falls back to v/v2/v3 instead of deriving real identifiers from the type. wire_gen.go is regenerated here to match the in-module build — a cosmetic rename of 9 lines, with no other change to the generated code.
golangci-lint deliberately keeps its action rather than becoming a tool: it pins its own x/tools (v0.44.0 vs our v0.48.0) for the analyzers it bundles, adding it to go.mod would double our go.mod/go.sum (158→338 / 441→889 lines), and the action supplies caching, only-new-issues and PR annotations that invoking a binary can't. Its version stays manual by request.
* Redact stream keys in UpdateStream API log fields
UpdateStream add/remove urls were logged raw into the API request log,
including rtmp stream keys and mux/twitch shorthand keys. Redact them
with utils.RedactStreamKey before appending log fields.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* Bump protocol for query-value redaction fallback
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Flush the downtrack sequencer on stream restart (Resync, ReceiverRestart,
codec change) so NACK retransmissions can't use metadata that no longer
matches the resynced bucket. Add a defensive bounds guard on the RTX and
forward payload slicing.
Cap the PacketHistory and FrameIntegrityChecker catch-up loops to the ring
size so a large sequence/frame-number jump can't drive a big per-packet
iteration count.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the per-NACK-packet goroutine spawn in DownTrack.handleRTCP with a
single long-lived worker that coalesces pending NACKs into one retransmit
pass. Pending sequence numbers are capped so a high NACK arrival rate cannot
grow goroutine count or memory unboundedly.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Bound the size of HTTP request bodies on the main API listener so large
messages cannot exhaust memory. Configurable via limit.max_api_request_body_size
(defaults to 10 MiB, 0 disables).
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Limit number of pending tracks per participant.
Prevents just a signalling connection adding tracks without actually
publishing them growing a large number.
* add to supervisor only if pending track is accepted
With https://github.com/livekit/livekit/pull/4706, there was a case of
some downstream component taking a long time while lock was held. While
the underlying cause of holding a lock while doing callback was removed
in that PR, to catch such cases, some publish side metric anomaly would
be useful to monitor and alert on.
Adding a publish time record for pending tracks on participant close.
That would inflate the publish time for participants not being able to
publish and can be alerted on as it will spike up the value at node
level and at cluster level if multiple nodes have the issue.
* Close web socket connections in all paths.
There was a leak of WebSocket pingWorker if the initial response write
errored as it did not close the WebSocket connection.
* graceful close
Shayne Parlo