_schemas.py was a local exception to the codebase convention. The rest
of the app keeps all API request/response DTOs in decnet/web/db/models.py
alongside UserResponse, DeployIniRequest, etc. — the swarm endpoints now
follow the same convention (SwarmEnrollRequest, SwarmHostView, etc).
Deletes decnet/web/router/swarm/_schemas.py.
Splits the three grouped router files into eight api_<verb>_<resource>.py
modules under decnet/web/router/swarm/ to match the convention used by
router/fleet/ and router/config/. Shared request/response models live in
_schemas.py. Keeps each endpoint easy to locate and modify without
stepping on siblings.
Adds decnet/web/swarm_api.py as an independent FastAPI app with routers
for host enrollment, deployment dispatch (sharding DecnetConfig across
enrolled workers via AgentClient), and active health probing. Runs as
its own uvicorn subprocess via 'decnet swarmctl', mirroring the isolation
pattern used by 'decnet api'. Also wires up 'decnet agent' CLI entry for
the worker side.
29 tests added under tests/swarm/test_swarm_api.py cover enrollment
(including bundle generation + duplicate rejection), host CRUD, sharding
correctness, non-swarm-mode rejection, teardown, and health probes with
a stubbed AgentClient.
- decnet.models.DeckyConfig grows an optional 'host_uuid' (the SwarmHost
that runs this decky). Defaults to None so legacy unihost state files
deserialize unchanged.
- decnet.agent.executor: replace non-existent config.name references
with config.mode / config.interface in logs and status payload.
- tests/swarm/test_state_schema.py covers legacy-dict roundtrip, field
default, and swarm-mode assignments.
decnet.swarm.client exposes:
- MasterIdentity / ensure_master_identity(): the master's own CA-signed
client bundle, issued once into ~/.decnet/ca/master/.
- AgentClient: async-context httpx wrapper that talks to a worker agent
over mTLS. health/status/deploy/teardown methods mirror the agent API.
SSL context is built from a bare ssl.SSLContext(PROTOCOL_TLS_CLIENT)
instead of httpx.create_ssl_context — the latter layers on default-CA
and purpose logic that broke private-CA mTLS. Server cert is pinned by
CA + chain, not DNS (workers enroll with arbitrary SANs).
tests/swarm/test_client_agent_roundtrip.py spins uvicorn in-process
with real certs on disk and verifies:
- A CA-signed master client passes health + status calls.
- An impostor whose cert comes from a different CA cannot connect.
Worker agent (decnet.agent):
- mTLS FastAPI service exposing /deploy, /teardown, /status, /health,
/mutate. uvicorn enforces CERT_REQUIRED with the DECNET CA pinned.
- executor.py offloads the blocking deployer onto asyncio.to_thread so
the event loop stays responsive.
- server.py refuses to start without an enrolled bundle in
~/.decnet/agent/ — unauthenticated agents are not a supported mode.
- docs/openapi disabled on the agent — narrow attack surface.
tests/test_base_repo.py: DummyRepo was missing get_attacker_artifacts
(pre-existing abstractmethod) and so could not be instantiated. Added
the stub + coverage for the new swarm CRUD surface on BaseRepository.
decnet.swarm.pki provides:
- generate_ca() / ensure_ca() — self-signed root, PKCS8 PEM, 4096-bit.
- issue_worker_cert() — per-worker keypair + cert signed by the CA with
serverAuth + clientAuth EKU so the same identity backs the agent's
HTTPS endpoint AND the syslog-over-TLS upstream.
- write_worker_bundle() / load_worker_bundle() — persist with 0600 on
private keys.
- fingerprint() — SHA-256 DER hex for master-side pinning.
tests/swarm/test_pki.py covers:
- CA idempotency on disk.
- Signed chain validates against CA subject.
- SAN population (DNS + IP).
- Bundle roundtrip with 0600 key perms.
- End-to-end mTLS handshake between two CA-issued peers.
- Cross-CA client rejection (handshake fails).
Introduces the master-side persistence layer for swarm mode:
- SwarmHost: enrolled worker metadata, cert fingerprint, heartbeat.
- DeckyShard: per-decky host assignment, state, last error.
Repo methods are added as default-raising on BaseRepository so unihost
deployments are untouched; SQLModelRepository implements them (shared
between the sqlite and mysql subclasses per the existing pattern).
decnet.collector.log / decnet.system.log and the *.db-shm / *.db-wal
sidecars produced by the sqlite WAL journal were slipping through the
existing rules. Extend the patterns so runtime state doesn't show up
in git status.
Reference template for .env / .env.local showing every variable that
decnet/env.py consumes, with short rationale per section (system
logging, embedded workers, profiling, API server, …). Copy to .env
and fill in secrets; .env itself stays gitignored.
Exercises the JSON → syslog formatter end to end: flat fields ride as
SD params, bulky nested metadata collapses into the meta_json_b64 blob,
and the event_type / hostname / service mapping lands in the right
RFC 5424 header slots.
Frontend now handles syslog lines from producers that don't use
structured-data (notably the SSH PROMPT_COMMAND hook, which emits
'CMD uid=0 user=root src=IP pwd=… cmd=…' as a plain logger message).
A new parseEventBody utility splits the body into head + key/value
pairs and preserves the final value verbatim so commands stay intact.
Dashboard and LiveLogs use this parser to render consistent pills
whether the structure came from SD params or from the MSG body.
The host-side sniffer interface depends on the deploy's driver choice
(--ipvlan flag). Instead of hardcoding HOST_MACVLAN_IFACE, probe both
names and pick whichever exists; warn and disable cleanly if neither
is present. Explicit DECNET_SNIFFER_IFACE still wins.
- Relaxed RFC 5424 regex to accept either NILVALUE or a numeric PROCID;
sshd / sudo go through rsyslog with their real PID, while
syslog_bridge emitters keep using '-'.
- Added a fallback pass that scans the MSG body for IP-shaped
key=value tokens. This rescues attacker attribution for plain logger
callers like the SSH PROMPT_COMMAND shim, which emits
'CMD … src=IP …' without SD-element params.
Each honeypot container now carries its own copy of the shared RFC 5424
formatter. Services that previously rolled their own ad-hoc syslog
lines can now import syslog_line / write_syslog_file for a consistent
SD-element format that the collector already knows how to parse.
Adds the server-side wiring and frontend UI to surface files captured
by the SSH honeypot for a given attacker.
- New repository method get_attacker_artifacts (abstract + SQLModel
impl) that joins the attacker's IP to `file_captured` log rows.
- New route GET /attackers/{uuid}/artifacts.
- New router /artifacts/{decky}/{service}/{stored_as} that streams a
quarantined file back to an authenticated viewer.
- AttackerDetail grows an ArtifactDrawer panel with per-file metadata
(sha256, size, orig_path) and a download action.
- ssh service fragment now sets NODE_NAME=decky_name so logs and the
host-side artifacts bind-mount share the same decky identifier.
The /opt/emit_capture.py, /opt/syslog_bridge.py, and
/usr/libexec/udev/journal-relay files were plaintext and world-readable
to any attacker root-shelled into the SSH honeypot — revealing the full
capture logic on a single cat.
Pack all three into /entrypoint.sh as XOR+gzip+base64 blobs at build
time (_build_stealth.py), then decode in-memory at container start and
exec the capture loop from a bash -c string. No .py files under /opt,
no journal-relay file under /usr/libexec/udev, no argv_zap name
anywhere. The LD_PRELOAD shim is installed as
/usr/lib/x86_64-linux-gnu/libudev-shared.so.1 — sits next to the real
libudev.so.1 and blends into the multiarch layout.
A 1-byte random XOR key is chosen at image build so a bare
'base64 -d | gunzip' probe on the visible entrypoint returns binary
noise instead of readable Python.
Docker-dependent tests live under tests/docker/ behind a new 'docker'
pytest marker (excluded from the default run, same pattern as fuzz /
live / bench).
The named pipe at /run/systemd/journal/syslog-relay had two problems
beyond its argv leak: any root-in-container process could (a) `cat`
the pipe and watch the live SIEM feed, and (b) write to it and inject
forged log lines. Since an attacker with a shell is already root
inside the honeypot, file permissions can't fix it.
Point rsyslog's auth/user actions directly at /proc/1/fd/1 — the
container-stdout fd Docker attached to PID 1 — and delete the
mkfifo + cat relay from the entrypoint. No pipe on disk, nothing to
read, nothing to inject, and one fewer cloaked process in `ps`.
Two leaks remained after the inotifywait argv fix:
1. The bash running journal-relay showed its argv[1] (the script path)
in /proc/PID/cmdline, producing a line like
'journal-relay /usr/libexec/udev/journal-relay'
Apply argv_zap.so to that bash too.
2. argv_zap previously hardcoded PR_SET_NAME to 'kmsg-watch', which was
wrong for any caller other than inotifywait. The comm name now comes
from ARGV_ZAP_COMM so each caller can pick its own (kmsg-watch for
inotifywait, journal-relay for the watcher bash).
3. The capture.sh header started with 'SSH honeypot file-catcher' —
fatal if an attacker runs 'cat' on it. Rewritten as a plausible
systemd-journal relay helper; stray 'attacker' / 'honeypot' words
in mid-script comments stripped too.
A lived-in Linux box ships with iputils-ping, ca-certificates, and nmap
available. Their absence is a cheap tell, and they're handy for letting
the attacker move laterally in ways we want to observe. iproute2 (ip a)
was already installed for attribution — noted here for completeness.
The kmsg-watch (inotifywait) process was the last honest giveaway in
`ps aux` — its watch paths and event flags betrayed the honeypot. The
argv_zap.so shim hooks __libc_start_main, heap-copies argv for the real
main, then memsets the contiguous argv[1..] region to NUL so the kernel's
cmdline reader returns just argv[0].
gcc is installed and purged in the same Docker layer to keep the image
slim. The shim also calls prctl(PR_SET_NAME) so /proc/self/comm mirrors
the argv[0] disguise.
exec -a replaces argv[0] so ps shows 'journal-relay /usr/libexec/udev/journal-relay'
instead of '/bin/bash /usr/libexec/udev/journal-relay' — no interpreter
hint on the watcher process.
inotify | while spawns a subshell for the tail of the pipeline, so
two bash processes (the script itself and the while-loop subshell)
showed up under /usr/libexec/udev/journal-relay in ps aux. Enable
lastpipe so the while loop runs in the main shell — ps now shows
one bash plus the inotify child, matching a simple udev helper.
Rename the container-side logging module decnet_logging → syslog_bridge
(canonical at templates/syslog_bridge.py, synced into each template by
the deployer). Drop the stale per-template copies; setuptools find was
picking them up anyway. Swap useradd/USER/chown "decnet" for "logrelay"
so no obvious token appears in the rendered container image.
Apply the same cloaking pattern to the telnet template that SSH got:
syslog pipe moves to /run/systemd/journal/syslog-relay and the relay
is cat'd via exec -a "systemd-journal-fwd". rsyslog.d conf rename
99-decnet.conf → 50-journal-forward.conf. SSH capture script:
/var/decnet/captured → /var/lib/systemd/coredump (real systemd path),
logger tag decnet-capture → systemd-journal. Compose volume updated
to match the new in-container quarantine path.
SD element ID shifts decnet@55555 → relay@55555; synced across
collector, parser, sniffer, prober, formatter, tests, and docs so the
host-side pipeline still matches what containers emit.
Rename the rsyslog→stdout pipe from /var/run/decnet-logs (dead giveaway)
to /run/systemd/journal/syslog-relay, and launch the relay via
exec -a "systemd-journal-fwd" so ps shows a plausible systemd forwarder
instead of a bare cat. Casual ps/ls inspection now shows nothing
with "decnet" in the name.
Old ps output was a dead giveaway: two "decnet-capture" bash procs
and a raw "inotifywait". Install script at /usr/libexec/udev/journal-relay
and invoke inotifywait through a /usr/libexec/udev/kmsg-watch symlink so
both now render as plausible udev/journal helpers under casual inspection.
fuser and /proc fd walks race scp/wget/sftp — by close_write the writer
has already closed the fd, so pid-chain attribution always resolved to
unknown for non-interactive drops. Fall back to the ss snapshot: one
established session → ss-only, multiple → ss-ambiguous (still record
src_ip from the first, analysts cross-check concurrent_sessions).
inotifywait watches writable paths in the SSH decky and mirrors any
file close_write/moved_to into a per-decky host-mounted quarantine dir.
Each artifact carries a .meta.json with attacker attribution resolved
by walking the writer PID's PPid chain to the sshd session leader,
then cross-referencing ss and utmp for source IP/user/login time.
Also emits an RFC 5424 syslog line per capture for SIEM correlation.
Commit-by-commit evidence of the perf work: each CSV is the raw
Locust output for the commit hash in its filename, plus the four
fb69a06 variants (single worker, tracing on/off, single-core pinned,
12 workers) referenced in the README baseline table.
Some pyinstrument frame trees contain branches where an identifier is
missing (typically at the very top or with certain async boundaries),
which crashed the aggregator with a KeyError mid-run. Short-circuit
on None frames and missing identifiers so a single ugly HTML no
longer kills the summary of the other few hundred.
asyncmy needs cryptography for caching_sha2_password (the MySQL 8
default auth plugin). Without it, connection handshake fails the
moment the server negotiates the modern plugin.
Capture Locust numbers from the fb69a06 branch across five
configurations so future regressions have something to measure against.
- 500u tracing-on single-worker: ~960 RPS / p99 2.9 s
- 1500u tracing-on single-worker: ~880 RPS / p99 9.5 s
- 1500u tracing-off single-worker: ~990 RPS / p99 8.4 s
- 1500u tracing-off pinned to one core: ~46 RPS / p99 122 s
- 1500u tracing-off 12 workers: ~1585 RPS / p99 4.2 s
Also note MySQL max_connections math (pool_size * max_overflow *
workers = 720) to explain why the default 151 needs bumping, and the
Python 3.14 GC segfault so nobody repeats that mistake.
Previous attempt (shield + sync invalidate fallback) didn't work
because shield only protects against cancellation from *other* tasks.
When the caller task itself is cancelled mid-query, its next await
re-raises CancelledError as soon as the shielded coroutine yields —
rollback inside session.close() never completes, the aiomysql
connection is orphaned, and the pool logs 'non-checked-in connection'
when GC finally reaches it.
Hand exception-path cleanup to loop.create_task() so the new task
isn't subject to the caller's pending cancellation. close() (and the
invalidate() fallback for a dead connection) runs to completion.
Success path is unchanged — still awaits close() inline so callers
see commit visibility and pool release before proceeding.
Under high-concurrency MySQL load, uvicorn cancels request tasks when
clients disconnect. If cancellation lands mid-query, session.close()
tries to ROLLBACK on a connection that aiomysql has already marked as
closed — raising InterfaceError("Cancelled during execution") and
leaving the connection checked-out until GC, which the pool then
warns about as a 'non-checked-in connection'.
The old fallback tried sync.rollback() + sync.close(), but those still
go through the async driver and fail the same way on a dead connection.
Replace them with session.sync_session.invalidate(), which just flips
the pool's internal record — no I/O, so it can't be cancelled — and
tells the pool to drop the connection immediately instead of waiting
for garbage collection.
Locust @task(2) hammers /auth/login in steady state on top of the
on_start burst. After caching the uuid-keyed user lookup and every
other read endpoint, login alone accounted for 47% of total
_execute at 500c/u — pure DB queueing on SELECT users WHERE
username=?.
5s TTL, positive hits only (misses bypass so a freshly-created
user can log in immediately). Password verify still runs against
the cached hash, so security is unchanged — the only staleness
window is: a changed password accepts the old password for up to
5s until invalidate_user_cache fires (it's called on every write).
The per-request SELECT users WHERE uuid=? in require_role was the
hidden tax behind every authed endpoint — it kept _execute at ~60%
across the profile even after the page caches landed. Even /health
(with its DB and Docker probes cached) was still 52% _execute from
this one query.
- dependencies.py: 10s TTL cache on get_user_by_uuid, well below JWT
expiry. invalidate_user_cache(uuid) is called on password change,
role change, and user delete.
- api_get_config.py: 5s TTL cache on the admin branch's list_users()
(previously fetched every /config call). Invalidated on user
create/update/delete.
- api_change_pass.py + api_manage_users.py: invalidation hooks on
all user-mutating endpoints.
Round-2 follow-up: profile at 500c/u showed _execute still dominating
the uncached read endpoints (/bounty 76%, /logs/histogram 73%,
/deckies 56%). Same router-level TTL pattern as /stats — 5s window,
asyncio.Lock to collapse concurrent calls into one DB hit.
- /bounty: cache default unfiltered page (limit=50, offset=0,
bounty_type=None, search=None). Filtered requests bypass.
- /logs/histogram: cache default (interval_minutes=15, no filters).
Filtered / non-default interval requests bypass.
- /deckies: cache full response (endpoint takes no params).
- /config: bump _STATE_TTL from 1.0 to 5.0 — admin writes are rare,
1s was too short for bursts to coalesce at high concurrency.
SQLite is a local file — a SELECT 1 per session checkout is pure
overhead. Env var DECNET_DB_POOL_PRE_PING stays for anyone running
on a network-mounted volume. MySQL backend keeps its current default.
Popen moved inside the try so a missing uvicorn falls through to the
existing error message instead of crashing the CLI. test_cli was still
patching the old subprocess.run entrypoint; switched both api command
tests to patch subprocess.Popen / os.killpg to match the current path.
Every /stats call ran SELECT count(*) FROM logs + SELECT count(DISTINCT
attacker_ip) FROM logs; every /logs and /attackers call ran an
unfiltered count for the paginator. At 500 concurrent users these
serialize through aiosqlite's worker threads and dominate wall time.
Cache at the router layer (repo stays dialect-agnostic):
- /stats response: 5s TTL
- /logs total (only when no filters): 2s TTL
- /attackers total (only when no filters): 2s TTL
Filtered paths bypass the cache. Pattern reused from api_get_config
and api_get_health (asyncio.Lock + time.monotonic window + lazy lock).
require_role._check previously chained from get_current_user, which
already loaded the user — then looked it up again. Inline the decode +
single user fetch + must_change_password + role check so every
authenticated request costs one SELECT users WHERE uuid=? instead of
two.
Only database, docker, and ingestion_worker now count as critical
(→ 503 unhealthy). attacker/sniffer/collector failures drop overall
status to degraded (still 200) so the dashboard doesn't panic when a
non-essential worker isn't running.
The ingester now accumulates up to DECNET_BATCH_SIZE rows (default 100)
or DECNET_BATCH_MAX_WAIT_MS (default 250ms) before flushing through
repo.add_logs — one transaction, one COMMIT per batch instead of per
row. Under attacker traffic this collapses N commits into ⌈N/100⌉ and
takes most of the SQLite writer-lock contention off the hot path.
Flush semantics are cancel-safe: _position only advances after a batch
commits successfully, and the flush helper bails without touching the
DB if the enclosing task is being cancelled (lifespan teardown).
Un-flushed lines stay in the file and are re-read on next startup.
Tests updated to assert on add_logs (bulk) instead of the per-row
add_log that the ingester no longer uses, plus a new test that 250
lines flush in ≤5 calls.
Adds BaseRepository.add_logs (default: loops add_log for backwards
compatibility) and a real single-session/single-commit implementation
on SQLModelRepository. Introduces DECNET_BATCH_SIZE (default 100) and
DECNET_BATCH_MAX_WAIT_MS (default 250) so the ingester can flush on
either a size or a time bound when it adopts the new method.
Ingester wiring is deferred to a later pass — the single-log path was
deadlocking tests when flushed during lifespan teardown, so this change
ships the DB primitive alone.
A module-level asyncio.Lock binds to the loop it was first awaited on.
Under pytest-anyio (and xdist) each test spins up a new loop; any later
test that hit /health or /config would wait on a lock owned by a dead
loop and the whole worker would hang.
Create the lock on first use and drop it in the test-reset helpers so a
fresh loop always gets a fresh lock.
Under CPU saturation the sync docker.from_env()/ping() calls could miss
their socket timeout, cache _docker_healthy=False, and return 503 for
the full 5s TTL window. Both calls now run on a thread so the event
loop keeps serving other requests while Docker is being probed.
With --workers > 1, SIGINT from the terminal raced uvicorn's supervisor:
some workers got signaled directly, the supervisor respawned them, and
the result behaved like a forkbomb. Start uvicorn in its own session and
signal the whole process group (SIGTERM → 10s grace → SIGKILL) when we
catch KeyboardInterrupt.
Forwards straight to uvicorn's --workers. Default stays at 1 so the
single-worker efficiency direction is preserved; raising it is available
for threat-actor load scenarios where the honeypot needs to soak real
attack traffic without queueing on one event loop.
