chunkloris: waitress

part of the chunkloris per-chunk amplification survey. this page is the per-server record for waitress under http/1.1 chunked transfer encoding.

at a glance

server: waitress 3.0.2
runtime: python-3.12
ecosystem: python
concurrency model: asyncore-main-thread-plus-worker-pool
parser: pure-Python waitress.receiver.ChunkedReceiver
delivery granularity: per-chunk-per-recv
chunk-limit helper: indirect-via-inbuf_overflow=524288
verdict: per-chunk — the parser/dispatcher boundary delivers one event per wire chunk. cpu cost under paced mode b is measurable per chunk.
scaling exponent (mode a): 1.45 (wall time vs N, log-log slope across common cells)

measurements

all cells run on a 1-vcpu docker container. cpu cost is derived from the target container’s cgroup v2 cpu.stat usage_usec delta around each cell.

mode	N	wall (s)	server cpu %	µs / chunk	basis	ok
`A-bridge-coalesced`	50,000	0.268	—	5.360	wall	✓
`A-bridge-coalesced`	100,000	1.720	—	17.200	wall	✓
`A-bridge-coalesced`	250,000	4.590	—	18.360	wall	✓
`B-paced-100us`	50,000	5.310	—	106.200	wall	✓
`B-paced-100us`	100,000	10.650	—	106.500	wall	✓
`B-paced-100us`	250,000	27.010	—	108.000	wall	✓

parser path — source citations

chunked-decoder — waitress/receiver.py:81 ChunkedReceiver.received → source
channel-handle_read — waitress/channel.py:154 → source
body-buffer-append — waitress/buffers.py:261 OverflowableBuffer.append → source

what this means

the parser/dispatcher path on this server delivers one event per chunked-transfer-encoding chunk, so an attacker who sends a body as N one-byte chunks consumes roughly N × (mode-b µs/chunk) of server cpu on a single core. amplification scales linearly with N until the framework’s max_request_body_size (or equivalent) is hit.

what to do today

if this server runs as an origin behind nginx with the default proxy_request_buffering on, the per-chunk attack shape does not reach this server — nginx delivers one content-length-framed body to the upstream in a single recv().
if deployed direct-exposed, behind haproxy with default streaming, or behind any reverse proxy with proxy_request_buffering off, the per-chunk cost reaches this server.
there is no framework-level chunk-count limit in the default config; use a frontend buffer, transport-layer rate limiting, or a wrapping middleware that imposes a chunk-count cap before draining the body.

reproducer

the full reproducer for this server is in the paper repo. the docker container pins waitress 3.0.2 and constrains the test container to a single cpu (--cpus=1). the prober script implements mode a (bridge-coalesced) and mode b (paced 100 µs) per the methodology section.

see the draft pdf for the full per-framework discussion.