The hello carries the fluksio-worker version. An agent refused for an older
protocol stays listed as Needs upgrade until it attaches again. Docs gain an
Upgrading workers section.
The connector now serves /mcp, its OAuth documents and /hooks/ besides
/api/v1/, refuses any dot segment (httpx resolves them, which carried
/api/v1/../../mcp onto the MCP mount), addresses calls to the hub's host,
and bridges served sockets by calling the app with a websocket scope.
A hook without a secret answers the tunnel with the 404 (or 1008) of a
wrong secret. Requests through the hub get the hub's issuer, resource and
consent page in the OAuth metadata and the 401, and the MCP host check
admits the hub while enrolled.
Every datetime the API emits for versions (flow summaries' updated_at,
runs' created/started/finished with flow_version) is already ISO 8601
with Z or +00:00. A test holds it and the API docs say so.
CACHE_SHAPE_SINCE marks the last change that moved every cache key. The
run service's periodic sweep clears cache_key on rows of runs created
before it, which nothing can match any more. Their outputs stay.
DELETE /runs?flow= removes every Run row of the flow but left its flow_run
rows, so a deleted flow kept appearing in the history. history=true now
drops those too, and the flow delete dialogs (editor and bulk) send it when
"Also delete its runs" is ticked.
Migration ccb21eacf5d5 deletes rows whose run is gone, then adds foreign keys
with ON DELETE CASCADE from run_node, run_metric and run_artifact to run.id.
Deleting runs is now the one statement. A node or metric written after its
run was deleted is dropped with an info line. Migrations run with foreign keys
off, since a batch rebuild of run would otherwise cascade into every child row.
The run detail works out curve_lost per node when read, from whether its
cached_from run still exists. The run page names the deleted run in place
of the generic NO_CURVE note.
Run.declared (migration 2e7ca7d10044) snapshots the declarations on submit
and import. The run detail's Inputs panel reads them, and a run from before
falls back to the flow's current declarations.
POST /observability/flush handles what the collector has queued and writes
it, so the soak harness no longer sleeps FLUSH_INTERVAL_S before clearing
its own rows.
A shared source goes live in every flow using it the moment it is saved, so
the write is recorded in the audit trail like a publish. Private source saves
stay drafts and are not audited.
DELETE /flows/library/{name} still answers 409 while nodes use the source.
With detach=true each of them keeps a copy of the code, in the published
flow and the draft, the source is deleted and the running users rebuild.
`ConnectorNode.store` keeps short strings per connector package in the
engine's database (`connector_state`, migration f8c3e5a91d27), so the nodes of
one package share them and `fluksio serve` needs nothing new: get, set and
delete, at most 64 kB a value and 256 keys a package. A superuser lists the
keys (never the values) and clears them through `/connectors/{node_type}/state`,
or under Stored state in a connector node's panel. Documented in the connector
contract.
The enrolment exemption covered any *.local host, which mDNS answers with
whoever claims it; it is now localhost, 127.0.0.0/8 and ::1. /docs, /redoc
and openapi.json stay closed in production and otherwise answer only a
loopback caller unless API_DOCS_ENABLED is set, which the dev compose does,
so `fluksio serve --host 0.0.0.0` no longer hands the schema to the network.
The limits read the leftmost X-Forwarded-For entry, which the client writes,
and the LAN nginx appended to it. The address now comes from uvicorn's
proxy-headers middleware (rightmost hop not in FORWARDED_ALLOW_IPS), applied
in the app too so tunnel-replayed calls resolve the same way; nginx sets the
header to $remote_addr and the compose backend trusts the Docker networks.
/users/signup gets a per-address limit of 10 an hour.
DELETE /flows/%2e%2e reached FlowStore.delete_flow(".."), which removed
the data directory, and /flows/_lib removed every shared node. Every
path the flow and dashboard stores build from a flow, node, library or
dashboard name now goes through checked(), which raises InvalidName
(400) for anything _validate_name refuses. _validate_name uses
fullmatch, so a trailing newline is refused too.
Writing node source, installing modules and any flow save that adds,
retypes or repoints a code node (Function, Command), edits a Command
node's settings or changes a node's resources.env now answer 403 for
everyone but a superuser. MCP tools and the assistant reach the same
routes, so they inherit the gate. Node types report runs_code.
The editor shows code read-only with a note, disables a Command node's
settings, leaves code nodes and reusable nodes out of the palette and
paste, and the Modules page offers no install.
/users/signup answers 403 unless SIGNUP_ENABLED is set. A public
GET /utils/config/ tells the login page, which hides the sign-up link,
and the signup page explains that an administrator creates accounts.
/search lists each input and batch output under its qualified message,
and picking one opens the flow on the label's id, which the editor now
treats like a click on the label: it opens the flow's settings.
`Run.passed` (migration d5a3c8e1f246) keeps the names the caller chose, so
the run detail and the comparison table can mark a value as the flow's own
again; rows from before it fall back to the names their params lack. Retries
hand on their parent's list, imports record their entry's.
Also: a cancel of a queued run no longer leaves its id in `_cancelled`, a
linked comparison no longer crashes before the list loads, and the pre-commit
failures are fixed (stale client, `text-[10px]`, worker formatting).
A claimed run only joined the leased set once its pipeline was built, so a
build waiting for a busy worker pool got the run swept as abandoned. Cancel
now also stops such a run, and a terminal status being overwritten is logged.
A history recorded before the move (MLflow experiments) had nowhere to
go, since runs were only made by running a flow. Each entry is stored as
it ended: parameters completed and checked like a submit, cause
"import", empty commit and code digest, metrics and artifacts written
as given, nothing queued or announced. Run.external_id (new migration)
makes a second import a no-op and is an export column so a row joins
back. The request is all or nothing and capped at 200 entries;
Client.import_runs pages a longer history.
`widgets.spec.ts` took a one-shot `boundingBox()` of a fill that is still
moving. Two ways that reads wrong. "a bar draws a row per reading" compared
two widths that are both zero until the socket has delivered the readings,
and a plain `expect` over a captured number does not retry. "a reading over
its scale fills the track and no more" allowed the fill one pixel past the
track, but the house spring is underdamped: `motion`'s own generator puts the
fluksio look at 100.66% of the track 276ms in (material 100.91%, glass
103.58%), which on this tile's ~240px track is ~1.6px — over the tolerance,
and reached only if the sample happens to land near the peak.
Neither is the widget. `fractionOf` clamps the reading to [0, 1] and
`.dui-bar-track` is `overflow: hidden`, so nothing past a full track is ever
drawn; a bounding box just is not clipped by an ancestor's overflow. So both
now poll: one until the shorter reading is drawn shorter, the other until the
fill is flush with its track — which is a stricter reading of "and no more"
than the pixel of slack was.
The live-socket test gains the other half of what that spin cost: an ASGI
task that never returns is one uvicorn's graceful shutdown waits on forever,
which is why `fluksio serve` could only be SIGKILLed and left `serve.pid`
behind.
`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.
`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 8c097d9 left behind when it corrected the
revoke_client docstring.
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.
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.
Several stateless instructions where a single button is one — six presets are
otherwise six tiles to place, six titles to read, and the message they share
repeated six times. Nothing is read back, as for a single button: what these
send is an instruction, and the last one sent is not a state to draw.
Auto-fit columns rather than a configured count: a tile is resized in the
editor and scaled again to whatever panel it hangs on, so how many fit is not
something the document can know.
The editor reads `cfg.buttons` raw rather than through `buttonsOf`, which
drops the blanks — a row being typed into is blank until the first keystroke.
Also carried along by the hooks: the generated SDK was stale (it had no
`RunsReadMetricNames`) and one pre-existing block in `test_panels.py` was
unformatted.
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.
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.
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.
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.
It described the wrong object. A dashboard is a document that may hang on a
hallway tablet and in a desk browser at the same time, and only one of those
has fingers on it — so the flag moves off `DashboardDef.settings` and onto
`PanelDef` as a plain bool, ticked in the Panels dialog. `useCanvasRoot` takes
it as an argument rather than reading the document, and `/panel/{id}` is the
only surface with a panel to ask.
Dropping the message binding with it is deliberate: nothing drove it, and a
flow deciding whether a screen has fingers on it was never the point. A stored
`settings.touch` is inert rather than migrated, which `_check_settings`
skipping unknown names already guaranteed.
The rail was the other half. It had no touch behaviour at all and its 40px
buttons met neither branch of the 44/32 rule. `[data-touch] .dui-rail{-item}`
in `ui/core/core.css` spends the padding and the gap on the buttons instead,
so they reach the 44px target and the rail comes out taller at exactly the
same width — `RAIL_INSET` never moves, and the arrangement under it does not
either.
Also closes the panels-dialog icon gap: `DashboardSummary` carries the `icon`
now, so the dialog draws each assigned dashboard's rail glyph beside its
checkbox. `initials()` went from three identical copies in the looks to one in
`Dashboard/icons.ts`, so the dialog and the rail fall back the same way.
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.
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.
`--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.
stroblme
Melvin Strobl