`asyncio.wait(..., FIRST_COMPLETED)` returns with at least one of the two
tasks done, so `receiver` missing from `done` already means `sender` is in
it. The `elif sender not in done: continue` could never fire.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CL9zvnnvcp1mvA8o7impxk
`receive_text` raises `WebSocketDisconnect` once the peer has gone and
`RuntimeError` on every call after that. The loop swallowed both — the
exception was checked only to decide whether the frame held a media
subscription — and immediately made another read, which raised again. A
closed tab therefore left a coroutine reading a dead socket at full tilt:
around 15,000 turns a second, one orphaned `Queue.get()` per turn parked on
the bus queue's waiter list. The event loop stalled behind it, `/health`
answered 503 on its loop-lag check, and Docker's probe plus Traefik took
the API out of the pool mid-request — which is what `make verify` was dying
on, with `TypeError: Failed to fetch` from the browser.
The spin ended only when an event finally reached the send path and the
socket raised there, so it lasted from milliseconds to minutes depending on
what the bus happened to be carrying. Four abandoned sockets logged 2,518
`Task was destroyed but it is pending!` lines and up to 2.6s of loop lag;
twenty now log none and cost 1.0ms.
A failed read is the disconnect, so it ends the stream. The getter created
beside it is cancelled rather than abandoned: nothing reads it after the
turn that made it, and left behind it would also take an event the live
reader wanted.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CL9zvnnvcp1mvA8o7impxk
`wanted_names` never read its `previous` argument — the caller reassigns
`wanted` from the return value — so the parameter goes rather than gaining
an underscore; its one call site and four test calls follow. Sorts the
imports in test_artifacts.py. Both were failing `make lint-backend` at HEAD.
Regenerates the client, which 8cb843e left behind when it corrected the
revoke_client docstring.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CL9zvnnvcp1mvA8o7impxk
Home's "recently modified" order was a proxy — drafts first, then the
version counter, then the name. Both summaries now carry `updated_at`,
read as the mtime of the working copy: every write in the store commits
immediately, so a file's mtime is its commit time, and `git log -1 --
<path>` costs ~990ms across this instance's 15 documents (it walks the
history back to the last commit touching each one, so it is slowest for
the stalest) against ~1.8ms for the stats. No cache needed.
`DashboardSummary` also carries `footprint`: each widget as its type plus
the placement a panel resolves, which is the whole of what the mosaic
draws. That drops the document fetch per tile and with it the cap of
eight, past which tiles showed a name and nothing else. Verified against
this instance's 8 dashboards: the blocks are identical to what the old
client-side derivation produced.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CL9zvnnvcp1mvA8o7impxk
An MCP access token is a stateless JWT good until it expires, so deleting
the client row revoked nothing already handed out — on the MCP endpoint or
on the REST API, which takes the same token directly. Both doors now look
the client up by the `client_id` the token has always carried, so tokens
already in circulation are held to it too.
A dashboard's `locked` setting stopped the client drawing a control and
nothing else; the server took a publish from a panel showing it anyway. It
now bounds the panel's write scope, resolved live where a flow drives the
flag, exactly as the client resolves it. Reads are untouched — read-only is
not blind — and so is a querying chart's request, which is how that tile
reads rather than something anyone touched.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CL9zvnnvcp1mvA8o7impxk
A run record is a record of what ran, so deleting a flow no longer sweeps its
history: `_forget_runs` is gone from the flow-delete path, and `DELETE /runs/{id}`
is the only thing that removes a run one at a time. The in-flight guard stays —
that is about work, not history.
`DELETE /runs?flow=` is the counterpart to the list's flow filter and what a
reseed needs, sharing the four statements with the single-run delete and
refusing the same way while a run of that flow is still going.
A run whose flow is gone reads as one: `useFlowInputs` reports the 404 rather
than an empty declaration set, so the run page says "flow deleted", explains it,
and disables Retry, which the route would refuse anyway.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CL9zvnnvcp1mvA8o7impxk
A python node's ports and settings arrive as keyword arguments, so a declared
name its `process` does not take was a TypeError on every call — and a node
that loads fine and fails every time it runs is the quiet kind of broken: the
hosted demo did it 720 times an hour for two days and the health badge read
ok throughout. `_build_node` now reads a written body with `ast` and refuses
the mismatch at load, so the node is an error on the canvas and an issue on
publish. Skipped for `**kwargs`, a decorated or absent `process`, and the
template a new node opens with. The SDK's generated shim always takes
`**settings`, so synced flows are untouched.
`/observability/summary` names a node that has failed in the last fifteen
minutes and reads degraded while it does, which is what would have made the
badge amber. `nodes.failing` carries the count.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SkgNtaR6JspnHBFFP6crZj
A share link admits somebody at the portal, so this instance first hears of
them when they arrive rather than when a superuser types their code in.
An unmapped portal identity is now checked once against the portal's own
list of who may reach this instance and given an ordinary local account
only if the portal vouches for it.
Asking rather than believing the token is the point: a token stays signed
and valid until it expires, so trusting its claims would let one rebuild
the account somebody deleted here and deleting a user would stop being the
whole of the revocation.
The account-making itself moved out of the route, since both ways in build
the same thing.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GPMNwB2mGBP5j7dXRopcPH
Each was a loose end recorded under `### SDK` in the notepad.
`serve` takes its own pidfile down on SIGTERM. uvicorn restores the handler it
found and re-raises the signal it stopped on, so the default handler ended the
process without unwinding and the `finally` never ran — which is what a stop
sends, and what left `serve.pid` behind.
`serve.log` is cut back past 5 MB by the engine rather than by the screen that
started it, so an adopted engine is bounded too. Gated on its own stdout being
an appended regular file, which is what makes the cut safe: the kernel then
puts the next write at the new end.
Cards are counted from `/dev/nvidia[0-9]*`, so `FLOW_GPUS`/`--gpus` of 0 means
"work it out" the way `FLOW_CPUS` always has. The engine counts, not the
accountant — a remote worker builds one of those from its own inventory, and
detecting there would hand it the engine host's cards. The worker counts last:
what a batch job says it was granted still wins.
`GET /runs/metrics/names` is the distinct over a selection that `--list` and
the terminal's metric picker were approximating by reading the newest run that
had measured anything, which missed a name only an older run ever wrote.
`MetricSink` announces each batch it has written (`run_metric`, carrying the
names). Not a per-point event: one covers up to 500 points or two seconds of
them, and the rows stay the record. The terminal comparison fills in as the
first readings land instead of staying blank until reopened, and the browser
refetches the run and any comparison rather than the list behind them.
`retry --group` pages the list route by `before` instead of stopping at 500.
The terminal dashboard takes the terminal's colours (`ansi-dark`), and the web
UI can re-pair from Settings without disconnecting first.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PRQ9bmTvCbqCwXo9mxZzzV
The site promises simulated inputs and mocked sensor values, and nothing in
the app was that. A run already is: the values are the caller's, the state is
the run's own namespace, and nothing it computes reaches the live flow. What
was missing was a screen to do it from, and the draft flag being honoured.
`/runs/new` is a flow, a field per declared input, a seed and Run; `/runs`
stays the log. A comma-separated list in a number field expands into the grid
`fluksio sweep --param` builds and goes to the sweep route, so launching one
no longer needs a terminal. Only numbers split: a comma in a string is
content, and one in JSON is syntax.
`RunCreate.draft` was validated at submit and dropped before the run
executed, so "try the working copy" ran the published one. `Run.draft` is a
column now, the driver reads the same copy the submit checked, and a retry
carries it. `FlowSummary.mode` came with it so the rail can say which flows
are batch before one is picked.
Also here: a Retry button on a finished run, which the route has always had
and the UI never did, and parameter cells truncated to their column with the
full value on hover.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013TTfoK82awm8wvxXhHz3XF
The rate the media dtypes could carry was one frame every second or two: each
was a file on the data volume, an event on the socket, and a request back for
the bytes. This closes both halves of that, and they are one feature.
`save_artifact(..., volatile=True)` writes to a `VolatileStore` — the same
content-addressed store, in `/dev/shm`, bounded by size with the oldest falling
out (`ARTIFACT_VOLATILE_BYTES`, 48 MB under the container's raised `shm_size`).
Nothing sweeps it: a frame nobody kept is not worth walking the store to find.
`ArtifactStore.path` falls through to it, which is what lets a volatile frame be
an ordinary reference everywhere else — the dtype check, a panel's digest scope,
`load_artifact` in a node, and the widget's own fetch all work on one unchanged.
`adopt` copies one into the store when a run records it, so "returned media is
kept, emitted media is not" stays true.
The bytes then go down the flows websocket as a length-prefixed binary frame,
sent just ahead of the `message_value` naming them, so a tile has the frame when
it hears the value moved. Nothing is pushed unasked: a client names the messages
it is drawing (`{"type":"media","names":[…]}`), a panel's list is intersected
with the scope it already had, and only the newest frame per name in a batch is
sent — a client that fell behind is not handed frames it would draw over. The
tunnel relays text only, so a screen reached through a portal falls back to
fetching, which is why the rate table now has two rows.
Around the edges: the remote worker's fetch cache is bounded at last
(`FLUKSIO_ARTIFACT_CACHE_BYTES`), since content addressing means nothing in it
ever expires and a media stream fills it with chunks nothing asks for twice; a
port carrying an image draws the frame in the node panel rather than only
saying `image/png · frame.png · 1.79kB`; and an edge chip says that much instead
of a line of hash. The media screenshot stops waiting for `networkidle` — a
camera is a socket that never goes quiet, which is the point of it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YC4u66vjzW54fnHu5Juhh9
serve: refuse a second engine for one data directory whatever port it was
asked for, using the pidfile and a token this directory signed. The check
runs before the database is touched and before the credential is written,
which is what left every later CLI call pointing at a dead port.
The terminal dashboard is three tabs (Overview, Runs, Logs) with the toolbar
following the focused pane, the engine's output goes to serve.log rather than
down a pipe, and closing the screen stops both reader threads so the prompt
comes back. It adopts a running engine on every start, so stop/start and
restart work on one it did not start, and a stop waits for the process to be
gone before the next start. Enrolment reports itself in the modal.
enroll: a new claim code replaces the pairing instead of being refused. The
code is redeemed before anything is written, mappings to a portal being left
are cleared, and a running engine redials when the stored enrolment changes.
runs: an engine re-queues the runs left `queued` by the one before it, and
`fluksio retry <id>` / `retry --group <sweep>` submits an interrupted run
again with the same inputs and group, recorded through Run.parent_id.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01U9BoNGq6V9MdRWAte7JBuC
Follows the portal: the noun is "instance" everywhere the app says it —
UI strings, CLI output, error details, docs and comments. The wire keys
(`instance_id`, `instance_token`) and the hub route this calls move with it.
An existing cloud.json is adopted rather than refused: without the key
alias the dataclass fails to parse, which the caller swallows and reads as
"never enrolled" instead of "reconnect".
`instance_key` on a node type becomes `target_key`. It means the outside
thing a node points at, which is a different sense of the word, and keeping
both would put two meanings of "instance" in one codebase.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015YrQnKV3bnQd4K342y8tKj
The log tee wrote through to the real stream unguarded, and the worker
pool tees a returned call's logs there after reading its result and
before handing it back — so a dead stdout, which `fluksio serve` makes
possible by running the engine as a child of the dashboard holding that
pipe, failed the node with its outputs already in hand. The capture half
runs first, so swallowing the write loses nothing.
Also: `flow_events` catches the RuntimeError a peer leaving mid-send
raises, which is a disconnect by another route, and the remote agent no
longer raises out of the task when its subprocess died before it could
be written to — the read below reports that and ends the call.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TXQv6KNyyvY7Z1etYTUUAd
`--format jsonl` existed, but `_cell` ran json.dumps at row-build time,
before a format was chosen — so a nested value was a string by then and
jsonl only re-escaped it, leaving a consumer against csv's 128KB field
limit either way. Stringifying moved to the csv writer, so csv is
byte-identical and jsonl carries json. The runs TUI followed, or a record
would draw as a Python repr.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KYM38KSb4V4v2T71eifnZv
A `webpush` alert channel, and the PWA it needs to arrive. The payload is
encrypted to the subscription (RFC 8291) and the request signed with this
installation's own keypair (RFC 8292), both over `http-ece` — `pywebpush`
does the same in one call but brings `requests` and `aiohttp` with it, two
HTTP stacks beside httpx on a machine that may be a Raspberry Pi.
The manifest and the worker are hand-written rather than `vite-plugin-pwa`:
there is nothing worth precaching when the page carrying the credential is
`no-store`, so the worker handles `push` and `notificationclick` and nothing
else. `registration.scope` is the app's root in both places it runs, which is
why the payload carries no URL.
A run finishing in error is the first event worth waking someone for; `ok`
and `cancelled` describe to nothing, so a nightly batch that works stays
quiet. The events were already on the bus — only the filter changed.
`WEBPUSH_FILE` is a derived path, so the keypair lands on the data volume
with the alerts beside it. Off it, a rebuild would silently stop every phone
being notified: the key they subscribed against would be gone.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014EbeFPm6WNC3YD9vrqqT3a
**SQLite is the database, and now says so.** `metric_minute` and every run
table are written with `sqlalchemy.dialects.sqlite.insert(...)
.on_conflict_do_update` and with `max(a, b)`, neither of which another
dialect has — so pointing `DATABASE_URL` at Postgres migrated cleanly,
served, logged in, and then lost every observability flush into the
collector's hold buffer and failed every run. It refuses at startup
instead. (The Postgres in the compose stack is Umami's; the engine's own
database has been a file beside the flows since 2026-08-21.)
**Every integer query parameter is bounded.** The caps were written as
`min(limit, 500)`, which a negative walks straight through — `?limit=-1`
compiles to `LIMIT -1` and SQLite returns the whole table. Ten signatures,
now `Query(ge=…, le=…)`. `hours=0` still means an hour, which
`_window_hours` was already deliberate about.
**Exports are capped at 10 000 runs** and say so with `X-Truncated`. The
filters bounded a sensible request and nothing bounded an unfiltered one,
which read every row into memory before a byte was streamed. `_series`
resolves cached curves in two queries rather than a `Run` lookup and a
`RunMetric` query per restored node — a comparison of twenty runs was
calling that twenty times over.
**`PUT /artifacts` has a size limit** (`MAX_ARTIFACT_BYTES`, 2 GiB, 0 to
disable), checked against `Content-Length` and again against the stream for
a chunked body, and its writes moved off the event loop.
**`/observability/timeseries` takes `since`/`until`**, the same window
`/runs` and `/events` take, capped at 2000 points — `hours=720&bucket_s=60`
was 43 200 of them in one array. It is also what a dragged chart needs to
re-fetch at its own resolution rather than magnifying buckets it has.
**Composite indexes** for the three list screens: `run(flow, created_at)`
and `(status, created_at)`, `flow_run(flow, started_at)`,
`engine_event(type, ts)`. Every index was single-column, so SQLite picked
one and sorted the rest by hand. Verified against a copy of a live database
(250k `flow_run` rows): the planner takes all four.
**Redis clients have socket timeouts.** A Redis that stops answering
without closing the connection hung the caller until the kernel gave up —
including `/utils/health/`, whose job is to notice.
**The panels file is written under one lock.** `save_panels` and
`unpair_panel` are both read-modify-write, and a save that read before an
unpair wrote put the old nonce back — silently un-revoking a screen that
had just been unpaired. The nonce carry-forward was written to make that
impossible; the gap between its read and its write is where it happened.
**Startup releases what it acquired.** Everything past `event_bus.bind`
registers how to close itself and the `finally` walks that list backwards;
a failure part-way through used to reach none of the shutdown steps and
leave the worker pool's subprocesses and every background task behind —
under `--reload`, once per bad edit. `modules.reconcile` moved into the
background: `uv` gets five minutes twice over, the healthcheck allows
eighty seconds, and the autoheal restarted the container before it could
finish installing.
`delete_run` takes SQLite's write lock up front (`core.db.writing`) rather
than upgrading a deferred transaction and losing to whichever flush
committed in between. `modules.sync` is serialised — two applies mutated
one venv at once. The proxied-call and stream dicts are bounded, and a
reused id cancels its predecessor instead of dropping the reference.
Security, found in passing and small enough to fix here:
- **`/secrets/` required only a signed-in user.** The names alone say what
this installation talks to, and `PUT /{name}` takes any name, so any
account could overwrite the credential a flow authenticates with.
Superuser now — which `/search` already assumed and said so.
- **`POST /login/access-token` had no rate limit.** Argon2 is deliberately
expensive and the route is unauthenticated and runs in the shared
threadpool. Ten *failed* attempts per address per five minutes; a
successful sign-in spends nothing.
- **a password reset link worked repeatedly for 48 hours.** The token now
carries a digest of the password hash it was minted against, so it stops
verifying once it has set one. No table of spent tokens needed.
- **enrolment accepted `http://`**, sending the claim code and then this
installation's credential in clear. https, or a local address.
- the rate limiter read `request.client.host`, which behind Traefik is the
proxy — so every per-address limit was one global bucket and one caller
could lock out everyone. It reads the forwarded address, and its
bucket table is capped rather than growing one key per address forever.
- SMTP has a timeout and sends after the response, so an unreachable mail
host cannot pin a threadpool worker, and a reply's timing no longer says
whether the address exists.
Test suite: engine-written rows are cleared between modules. A `FlowRun`
left `running` by one module turned up in another's query. Per-test
rollback is not available here — the module-scoped `client` runs the real
lifespan and its collector and run service write through sessions of their
own — so this bounds it where the writes come from. Three consecutive
green runs, orders randomised.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M6hPWS6YEbT1P8LxhhFb2T
A three-node cascade publishes 13-16 events and each one crossed to the
event loop on its own. They are one `call_soon_threadsafe` now — whatever
was published between two turns of the loop goes over together — and every
subscriber still receives every event, oldest still dropped first when one
falls behind.
The socket end of the same path:
- **any frame from the client ended its stream.** `receive_text` was
awaited once, outside the loop, so a keepalive — or anything else a
client decided to say — satisfied it and was read as the client going
away. It is recreated per iteration; only a disconnect ends the stream.
- events go out in one frame per wave (`{"type": "batch", "events": [...]}`,
capped at 64), serialised once with orjson rather than per client with
the stdlib's `json.dumps` through `send_json`. The client unpacks a batch
and still understands single frames, so an older engine behind a newer
bundle keeps working.
- authenticating and building the snapshot happen on a thread. Both were on
the event loop: one is a database round trip, the other reads the whole
of state, per connect and again per `dashboard_changed` per panel.
`Pipeline.values()` — what that snapshot is — no longer SCANs the whole
Redis namespace. It scanned five bookkeeping keys for every message to find
the messages; `RedisState` keeps a set of the names beside them and answers
from it. Maintained wherever a message is written, so a seeded value or a
deleted flow keeps it exact.
On the client, while in the same file:
- a `node_health` event invalidates the flow's detail. The canvas draws
health from the server-derived `issues`, so a node going down or
recovering only showed on mount, navigation or a rebuild. The store had
a health map of its own that nothing ever read; it and `useNodeHealth`
are gone rather than wired up, since the server's view is the one the
canvas already uses.
- a reconnect invalidates the five key families this socket feeds instead
of the entire cache, and the backoff is jittered. The usual reason a
socket dropped is the engine restarting, so every tab and every wall
panel refetched everything, together, at the moment it was least able to
answer.
- a frame that will not parse costs the frame, not the connection. It was
the one unguarded `JSON.parse` in the app; an exception there escaped to
`window.onerror` and left whatever it had already applied behind.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M6hPWS6YEbT1P8LxhhFb2T
`save_artifact` had no download counterpart: the run detail listed a run's
files and nothing in `fluksio --help` fetched one. `fluksio artifacts RUN`
lists them, `fluksio artifacts RUN NAME` writes one — under the name the
node saved it as, since the message name is chosen for the graph. The run
detail now carries that filename, which it held in the table and did not
report.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019Hra4ndWMCLU5F3KjUuVAc
`export runs` dropped a `param.*` column whose value was constant across the
exported runs, so a downstream filter broke depending on which runs the
selection happened to hold. Every input the selection recorded is a column
now; `--params` still narrows it to a sweep's axis. The metrics default is
unchanged.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019Hra4ndWMCLU5F3KjUuVAc
Runs: a run can now be deleted (DELETE /runs/{id}, cancelling a live one
first), exported as csv from the screen's own filters, and its flow label
opens the flow. Its "Parameters" panel became "Inputs" and lists every
input the flow declares, marking the ones that took the flow's own value
rather than the run's — the comparison table resolves the same defaults
instead of printing "unset".
Home reads brain, dashboards, health, flows: the mosaic is one full-width
scrolling strip, and the flows list and the flow-activity rollups merged
into a single left-joined table so a flow's state and its numbers sit on
one row.
Charts take a drag to narrow the x window and a double click or tap to
come back out. UplotChart holds the scale and passes resetScales:false
while a window is held, which is what the old comment said made this
impossible.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019LrWVRguqbk33YzfEeUx5W
`GET /api/v1/search/` hands the client one flat index of everything worth
jumping to — flows and the nodes inside them, dashboards and the widgets on
them, panels, secrets, modules, workers and alert channels — and cmdk matches
it in the browser, so results narrow while typing without a round trip per
keystroke. A node hit is the one thing no list endpoint could answer: it opens
its flow with that node in focus.
The panel is reached from **Search** above Documentation in the sidebar, or
⌘K anywhere. The flow canvas palette moves to ⌘P, being the narrower of the two.
The panels dialog gains an address (`/dashboards?panels`) so a panel hit has
somewhere to land, and the sidebar logo gets `shrink-0`: the rail's width
animates while the logo is already back, and a flex item short of room is
squeezed rather than clipped.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016vGH7jqcXxWKP9wZFPyVdU
The poll loop remembered what it read rather than what it published, so a
value the node could not publish counted as said: the next poll skipped it,
succeeded, and health went back to ok with the port still dark. Remember it
only after inject returns, and report ok last.
A node reporting itself down is now derived into its flow's issues on read
and counted on the health summary, so the canvas marks it and Home says so.
Being down does not stop the flow, and the issue clears by itself when the
node reports well again. The repeating poll warning is logged once per
outage rather than once per tick.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K1moruzue2kTJd3uVisgNk
Three things the first export pass got wrong for a real study.
**Dotted paths.** A node returns a record, not a scalar — the numbers arrive
inside `final_metrics` — so `--metrics final_metrics.train_loss` yielded an
empty column and `--metrics final_metrics` yielded the whole record in one
cell. Both sides of the wide table now take dotted paths, and the defaults
reach the same depth: every number a result carries is a column named by its
path, and inputs are compared leaf by leaf, so two configurations differing in
one field give that field as the axis rather than two blobs that are merely
not equal. Lists stay whole — a curve belongs in the long table.
**`--list`.** Metric names are flow-qualified, so `--name train_loss` matched
nothing and said only that. `fluksio export metrics --list` prints the names
the selection carries, and an empty export made with `--name` points at it.
**A version to compare.** The CLI ships ahead of the engine and a stale one
answered a flat 404 with nothing anywhere in the API to tell how old it was.
The engine reports `version` on `/observability/summary`, `fluksio status`
prints it, and a 404 from export now names both versions — or says "older"
when the field itself predates the engine. Bumped to 0.1.5, which is what
makes the number worth reading.
Also formats `flow/metrics.py`, which had been committed unformatted and was
the last `ruff format --check` failure.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A9Hdrmf2cwNABCnE5x9UJa
`fluksio export metrics` is the long table — a row per run, metric and step —
and `fluksio export runs` the wide one, a row per run with the inputs that
*vary* across the selection as columns beside its final numbers, status,
duration and the commit and digest of the code it ran. Both carry the run id
on every row, which is the join back to the run page and what makes an
exported file auditable. `Client.export_metrics`/`export_runs` answer the same
rows to a notebook.
The engine streams csv or jsonl from two routes declared above `/{run_id}`;
parquet is a client-side conversion behind the new `fluksio[parquet]` extra,
so nobody pays for pyarrow who does not want dtypes kept. The long export
reads each run through `_series`, so a cached node's curve comes with it, and
`--stride` thins each series rather than the concatenation of all of them.
Two things they needed on the way: `GET /runs` takes `?since=` and `?before=`,
so a long history pages by the last row's own timestamp instead of an offset
that shifts under it; and a read that reaches no engine now says so in half a
second rather than seven, because `runs`, `flavors`, `export` and an unwatched
`status` pass `retries=0`. Everything that submits keeps them.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A9Hdrmf2cwNABCnE5x9UJa
A bool was excluded from the history as "not a measurement", so a true/false
port had no curve in the node panel and none on an edge — only the word. It is
recorded as 0/1 now and drawn as steps, since a bezier through two states
slopes through readings that never happened. The axis is pinned to 0..1, so a
flag that was never on sits at the floor rather than mid-box.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The newest metric bucket is upserted every flush while its minute runs, and
both rollup endpoints summed it in — so every curve on Home ended on a fall
that was only the clock. The timeseries now stops at the last closed bin, and
the flow rollups' window ends on the last closed minute, so all sixty slices
are whole ones.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Concurrent runs sat at 4 whatever FLOW_MAX_CASCADES said: that setting bounds
cascades, and the run drivers read a hardcoded MAX_PARALLEL nobody could reach.
FLOW_MAX_RUNS is the knob they read now, --max-runs/--max-cascades/--max-workers
are the same three as flags on serve, and the engine says which numbers it
started with — which is the only way to tell that a settings file was read.
Events keep the run they happened in. The payload always carried it and the
persist path dropped it, so reading one run's failures meant filtering the
engine-wide list; a batch run's id reaches those events now too, since a run
has no journaled item to name itself by.
Also: a provisioner's 0 means "no deadline" rather than "cancel on the next
reconcile", and a command that reaches no engine says how to start one.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015sbYeYaVgYQqm1sbx7wPdL
Raw cpus and gpus are a property of the machines an installation has, so a node
written against a cluster quietly stops meaning anything when the cluster is
replaced. A node says "gpu-small" instead, and what that is stored here —
editable, and read again every time the node is built, so changing the flavor
changes what the next run gets.
Memory joins the schema properly (`ram`, in MB, accepting "2G"), along with
`duration_s` for how long a node is expected to take. That one is recorded and
shown and nothing else yet: a statement for whoever is planning around the node,
not a limit — the limit is still `timeout`.
A flavor and a number for the same thing is refused, compared by value so an
editor writing the whole object back with its defaults still round-trips. A name
nothing stores is refused at the save, which covers the canvas and `fluksio
sync` at once, and deleting one a node still asks for says which node.
Four sizes are seeded on an installation that has none, and never re-seeded:
re-adding one somebody deliberately removed is an argument nobody wins.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6HeySA27EkGANZN95QySW
The engine answered "where does this node run" twice, in two ways that could
not see each other: a device sent it to a worker carrying that label, and
resources were counted against the engine's own cores. Declaring both meant the
second answer won and nothing was counted at all — which the data-science
getting-started page and the worked example both do.
One question now, in flow/placement.py: of every machine attached, which could
grant what this node asked for, and which of those has it free. The books move
onto each machine — one accountant per worker, built from the inventory it
reported — and the waiting moves above them, where one condition variable can
be woken by a release anywhere or by a worker attaching. Locks go one way:
placer, then a machine's books, never back.
So a node asking for a card now finds the box that has one, rather than being
clamped down to none and run here. When nothing can grant the ask at all it is
still cut down and run — a flow written on a cluster has to work on a laptop —
but the ceiling is one real machine now, since taking the largest of each
dimension separately can describe a machine nobody has.
Two things fixed on the way. A device on a connector node held every batch run
of its flow forever, waiting for a worker that could never run an entry point.
And `prefer` falling back to the engine skipped the books, so the fallback held
nothing.
The bench flow's node has taken a `params` argument that with_settings has not
forwarded for some time, so the benchmark could not run at all: 62 ms median
submit-to-result with this, against the 61 ms on record.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6HeySA27EkGANZN95QySW
A worker reported its labels and nothing about the machine behind them, so the
engine could route a node to a GPU box but not tell whether that box had a GPU
free. Inventory — cores, GPUs, memory — now arrives with the hello frame, and
the run frame carries back what the engine allocated for that call.
Which is protocol 2 on both ends. GPUs are never probed: asking a vendor tool
would make the one dependency two, so a GPU is what the batch job says it was
given or what --gpus says. A worker that reports nothing still attaches and is
scheduled by its label alone.
Two things a job scheduler needs: --max-idle stops a worker started for one job
rather than letting it hold its allocation to the walltime, and a refusal is now
fatal instead of a reconnect loop that reads as a hang in a job's log.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6HeySA27EkGANZN95QySW
A port may now declare `image`, `audio` or `video`. Each is the artifact
reference the engine already had, narrowed by the `media_type` on it, so a
speech recogniser declares what it eats rather than taking any bytes at all and
finding out. Bytes still never travel as a message and nothing on the wire
stops being JSON: a camera publishes one reference per frame, a microphone one
per chunk, and a reference may carry a `meta` dict nothing here interprets.
Streaming media is therefore an ordinary streaming port — with one change to
what that means. An emission used to journal an item with no payload, so
downstream read whatever was current when the item was claimed; a consumer
slower than its producer saw only the newest chunk and the ones between were
lost. That is right for a training curve and wrong for a second of speech, so
an emission now journals a `kind="emission"` item carrying its values, and the
executor hands them to the nodes reading that message instead of writing them
to state again. The value in state stays the latest, which is what everything
else reads, and the wave is filtered by what actually changed rather than
walking everything reachable. No queue serialization change — the existing
`outputs` field carries it.
Continuous media makes the store's missing GC a real problem, so this closes
it: `sweep_artifacts` runs hourly, keeps every digest a `run_artifact` row
records or a live message holds, spares anything written in the last hour, and
stands aside entirely while a run is in flight, since a node may store a
checkpoint long before it returns the reference to it. That also collects the
orphans a deleted flow has always left behind. `ARTIFACT_GC_INTERVAL_S=0` turns
it off.
Around the edges: `GET /artifacts/{digest}` serves the media type the caller
passes and answers ranged requests, so a browser plays a clip rather than
downloading it; `PUT` spools to disk instead of holding the whole body in
memory, as does `save_artifact` given a path; a Media widget draws whatever its
message points at, and a wall panel may fetch the bytes its own tiles are
showing and nothing else; and a connector gets `save_artifact`, for a device
whose readings are bytes.
What this cannot do is live video: a frame every second or two is a glance, and
the honest answer above that is the camera's own stream, which the widget takes
as a URL and the browser plays from source.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Five concurrent training nodes, each sizing its thread pool to every core,
left the engine's own event loop unscheduled: the API stopped answering
within 10 s and every client died. The same shape on a GPU deadlocked a run
for 21 minutes at 0% utilisation with nothing failing and nothing to read --
it just sat in `running`.
@node(resources={"cpus": 2}) is the declaration. The engine holds that much
for the length of the execution, so more of them than the machine has room
for wait their turn rather than oversubscribing it, and a `gpus` node holds
its card exclusively. FLOW_CPUS defaults to every core but two, and those two
are what keeps the engine answering.
Because a thread cap is read when the process imports the library, a warm
worker cannot be told a different one -- so an environment gets a pool of its
own and nodes deriving the same one share it, rather than paying a cold start
per call on exactly the nodes whose imports are slowest. XLA_FLAGS is never
derived: it is a composed, version-dependent string, so it travels in
resources.env where it is visible.
A node that declares nothing is not accounted for and behaves as it always
did -- it just gets FLOW_CPUS/FLOW_MAX_WORKERS as a thread cap, which is the
half of this that fixes the reported incident without anybody declaring
anything. An operator who set OMP_NUM_THREADS themselves still wins.
Resources are claimed strictly before a worker slot, so the two blocking
waits cannot deadlock. A node queued for them publishes node_queued and shows
on GET /workers/resources, because waiting and hanging looked identical.
Accounted, not enforced: no cgroups, no rlimits. Scheduling across machines,
flavours and enforcement are the next steps.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Three faults with one root: the stored body of a code-defined node is an
import shim, and nothing that mattered was ever read from the code itself.
- The run stamp could not identify what ran. The shim imports whatever is on
disk when the worker starts, and an uncommitted tree stamps <commit>-dirty
for every run it ever produces. Run.code_digest hashes the repository's .py
files, memoized on their stat state, and it is read again when the run is
actually claimed -- so a sweep queued for hours records the code each of its
runs executed, not the code that was there when it was submitted.
- The stage cache adopted code that was too new. The fingerprint hashed the
shim, which is invariant under any edit to the imported function or anything
it calls into, so a re-run was served from cache and answered without the
outputs the edit added. It now carries the repo digest and the node's
declared ports. Every fingerprint changes once, which invalidates the
existing cache; a canvas flow has no repository and keys as before.
- An interrupted sync looked like a hand-edited canvas. The engine answers a
new-node template for a node with no stored body, and the template carries
no marker, so the drift check read "somebody edited this" and demanded
--force -- for the one state that re-running the sync is the fix for.
NodeSource.missing states the fact, and sync skips those and reuses the
bodies it read instead of asking for each one twice.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A driver script died of one slow answer: httpx.ReadTimeout out of
RunHandle.refresh() with a 30 s read timeout and no retry anywhere, which
cost a sweep 78 of its 84 runs.
- Split the timeout (5 s connect, 120 s read): a wrong URL fails at once,
and a busy engine gets longer than the slowest thing it does on purpose
(a 60 s compile, a 15 s rebuild wait).
- Retry idempotent calls three times on a transport error or 502/503/504.
503 is the engine's own "ask again" — it is what RebuildBusy answers.
- Submit carries a key the engine stores with the run, so a retry after a
timeout returns that run instead of starting a second. A sweep keys every
entry, so a half-created one recreates only what is missing.
- wait() and --follow tolerate five failed polls in a row; a 404 still stops
at once, because that is an answer rather than a gap.
- CLI says "engine not answering" and names the run still on the engine,
instead of printing a traceback.
- runs: clamp the params column to 80 characters; events() takes the
flow/since/until the endpoint already had; RunHandle.failures answers
"what killed this run" from the run's own node rows.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`RedisWorkQueue.stats` read XPENDING, which counts entries delivered to a
consumer and not yet acknowledged — work in progress. Entries sitting in the
stream undelivered were counted nowhere, so an engine hours behind reported
itself idle: on the house, `pending: 4` while the group's lag was 1554.
The group's own `lag` is the missing number. `backlog` now carries it on both
queues (`len(_items)` in memory), leads the health tile, and a sustained one
publishes `engine_degraded` from the timer thread — named with the flow most
of the waiting work belongs to, sampled from the undelivered tail, since that
is the actionable half. It is a summary problem rather than a /utils/health
503: a backlog should not restart the container.
Also drops the keyspace `scan_iter` `stats()` did per poll to count parked
items — it walked every state and idempotency key twice per ten seconds — for
a set the park/unpark path maintains.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpfSinyCBfjuieikyfMPbf
Clearing a node's last error on the engine published nothing, so a second
browser kept the marker until its next snapshot. One event carries the
qualified node; the receiving client drops the marker without refetching.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013Gf7WaExcJ9bs3kfJXB3nK
`POST /runs/flows/{name}` hardcoded `cause: "api"`, so every row in the
history claimed the same origin. The body now carries an optional `cause`,
closed to the values the column knows — the dashboard sends nothing and stays
"api", `fluksio run` says "cli", and the SDK client says "sdk".
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013Gf7WaExcJ9bs3kfJXB3nK
`ADVISORY_ISSUES` moves next to `ValidationIssue` in pipeline.py, and the
model derives an `advisory` flag from its own code, so the distinction the
engine already made ships to the client instead of being re-guessed there.
The dock keeps its summary in `--destructive` only when a real fault is
among the issues and paints an advisory row `--muted-foreground`; the
canvas leaves advisories off a node's dot and border entirely, since node
status has three colours and no warning tier.
biome checks the generated `openapi.json`, which nothing formats since the
SDK script dropped its format pass — ignore it like the other generated
files.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013Gf7WaExcJ9bs3kfJXB3nK
A cache hit still replays no emissions — those values were the story of an
execution that is not happening — but the run they were recorded in is now
written on the row (`run_node.cached_from`), and the metrics endpoints read the
series back from there. So a reused run answers `run.metrics("train.loss")`
with the same points the run that trained did, rather than looking like a run
that produced no numbers at all. Pointed at rather than copied: a sweep of 500
reusing one frozen node would otherwise duplicate its curve 500 times.
That needed the cross-flow restore fixed first. The cache key has no flow in
it while the stored outputs are named for the flow that produced them, so
`quick.prepare` getting a hit from `train` wrote `train.dataset` into `quick`'s
state and the next node was called without its argument. One rule now covers
both halves: `requalify` reads a name owned by one flow as the same name in
another, applied to the restored outputs, to the node id behind the pointer,
and to the series names on the way out. Reuse across flows is kept.
Also: `@run:<id>.<output>` and a bare `sha256:` digest resolve on every input,
not only artifacts. Chaining a run's json config into the next one from a shell
meant pasting the whole object inline, and the CLI could not even send the
spelling — `_coerce` died in `json.loads` before the engine saw it. Both
spellings are reserved on every input now, `str` included, and `_from_run`
returns whatever the run's result holds rather than only a reference.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dp9L6gakMVro1K2C5zdtBE
Four things from a testing pass.
`fluksio serve` printed its own lines through the root logger, which has no
handler and falls back to `INFO:fluksio.cloud.connector:...` — beside uvicorn's
aligned output it reads like something went wrong. The engine's loggers and
alembic's now use uvicorn's own handler. Named rather than configuring the
root: httpx logs every portal call at INFO and none of that is printed today.
`fluksio enroll` writes its config from another process, so an engine already
serving never learned it had been paired. It now looks for one every few
seconds and dials when it appears. `load()` rather than `exists()`, or a file
that does not parse would be restarted forever.
`fluksio status` says where the installation stands with its portal — never
paired, linked, or paired and unreachable, which is the one worth acting on.
`--seed` and `--timeout` had no help text at all. Both say what they are for
now, and the docs say what a seed is actually for: recorded on the run, part of
its input digest, and passed to an input named `seed` when the flow declares
one, so the number a run is labelled with is the one the code drew from.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019V5bsYGNxcgPs4xXmTPx69
A code node in a batch run is now fingerprinted by its source, its raw
settings and the values it reads — an artifact input counting as its digest,
which is what the content addressing was always for. A run that finds the key
restores what the earlier one returned and skips the node, recorded as
`cached`. The run history is the cache: `run_node.outputs` beside the
`cache_key` the schema already had, no second store. On for code nodes, never
for the built-in and connector types that have side effects; off per node with
`@node(cache=False)` and per run with `--no-cache`.
Emissions are not replayed on a hit, so a cached training node returns its
result without redrawing its curve. Recorded in NOTEPAD.md with the two other
deliberate limits.
`fluksio run --local` boots the real app in the command's own process and
drives it through its ASGI interface behind the ordinary client, so a run no
longer needs a `serve` terminal beside it — same data directory, same history,
and the cache carries between the two. It always waits, because the engine it
starts lives exactly as long as the command.
Also: `fluksio sweep --param lr=0.1,0.01` for the product of the lists,
`run --follow` for a run's numbers as they arrive, Ctrl-C cancelling a waited
run rather than abandoning it, coloured statuses on a terminal, and `name`
made optional on the metrics endpoint so a follower can ask for every series.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The gates have never gone green on the new runners. Three separate reasons:
- backend/Dockerfile shipped Python 3.10 while the code imports typing.Self
and datetime.UTC, so the container exited on import and the suite could not
even load its conftest. The image moves to 3.13 and the packages declare
>=3.12, which is the floor the tests actually pass on; ruff's target follows
and rewrites timezone.utc and asyncio.TimeoutError accordingly. Relocking
drops the 3.10 branch, which bumps FastAPI and so regenerates the SDK.
- frontend/README.md had no trailing newline and two dashboard widgets used
arbitrary text-[…] sizes. Both are em-relative on purpose, so they move to
the inline style the neighbouring ramp already uses.
- Every commit left its own run queued: without a concurrency group a runner
that was offline for a while works through a backlog nobody reads. A stack
that fails to come up now prints its logs before the teardown removes it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A data scientist keeps their code where it is and decorates it: `@node`
declares a function's ports beside the function, `Flow(name, nodes=[...])`
says which of them make a flow, and `use(fn, wire=..., **settings)` rebinds
one for a single flow. `fluksio sync` uploads the document plus a generated
import shim per node, so the store still holds a complete, runnable,
git-versioned definition while the code it imports stays theirs.
`fluksio login|run|runs` and `flow.submit().wait()` are the client half, over
the run endpoints that already existed. Runs record the user repository's
commit beside the store's, so "what code produced this number" is answerable
on the side that now holds the code.
- `fluksio/sdk/`: ports, decorators, the flow builder and its checks, the shim
generator, an HTTP client and sync. Standard library only at import, so
`from fluksio import node` in a training script pulls in no engine.
- `FlowDef.origin` marks a flow code-defined; `Run.origin_commit` carries the
repository's commit; `POST /modules/refresh` retires the workers without an
install, which every sync calls — a worker holds the imported package in
memory, so an edit to it is invisible until the process goes.
- The canvas shows a generated body read-only and names the repository to edit
instead; a body edited there stops the next sync rather than being discarded.
- The worker's reporter carries inert `Port`, `node`, `use` and `Flow`, since
the shim imports a module whose first line declares them.
- `examples/myresearch` is the worked example, `make sync-example` uploads it.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012ue1tkFWB1bcGy3aWhCKpU
stroblmeandClaude Opus 5