Protect task diagnostics and stream incremental run output
Build custom container images / build (map[base_image:php:8-fpm-alpine build_args:PHP_VERSION=8 context:php8-pgsql fingerprint_command:{ apk info -v | LC_ALL=C sort; find /usr/local/lib/php/extensions /usr/local/etc/php/conf.d -type f -exec sha256sum {} + | LC_ALL=C sort; } name:ph… (push) Successful in 38s
Build custom container images / build (map[base_image:postgres:18 build_args:PG_VERSION=18 POSTGIS_VERSION=3 VCHORD_VERSION=0.5.3 context:postgres fingerprint_command:{ dpkg-query -W -f='${binary:Package}=${Version}\n' | LC_ALL=C sort; find /usr/lib/postgresql -type f -exec sha256su… (push) Successful in 47s
Build custom container images / build (map[base_image:python:3 build_args:PYTHON_VERSION=3 context:python-tools fingerprint_command:{ dpkg-query -W -f='${binary:Package}=${Version}\n' | LC_ALL=C sort; pip freeze | LC_ALL=C sort; } name:python-tools oci_labels:org.opencontainers.ima… (push) Successful in 1m24s
Build custom container images / build (map[base_image:python:3.12-slim build_args:PYTHON_VERSION=3.12-slim context:linkki-tiedotus fingerprint_command:{ dpkg-query -W -f='${binary:Package}=${Version}\n' | LC_ALL=C sort; pip freeze | LC_ALL=C sort; } name:linkki-tiedotus oci_labels:… (push) Successful in 38s
Build custom container images / build (map[base_image:python:3.12-slim build_args:PYTHON_VERSION=3.12-slim context:timepoll fingerprint_command:{ dpkg-query -W -f='${binary:Package}=${Version}\n' | LC_ALL=C sort; pip freeze | LC_ALL=C sort; } name:timepoll oci_labels:org.opencontai… (push) Successful in 43s

This commit is contained in:
ajp_anton committed 2026-10-26 23:39:50 +00:00
1 parent 0aa67439fd
commit 55bbbe2fd7
8 files changed
+315 -47

No files matched your search

+12 -6
View File
@@ -8,9 +8,9 @@ with `python -m app`.
The application discovers direct `*.py` files in each non-hidden directory
under `/opt/tasks`. Each directory is a task group. Tasks in one group run
sequentially; different groups may run concurrently. It records run history and
combined output in SQLite under `/var/lib/server-maintenance`, and supports
cancellation of active task process groups.
sequentially; different groups may run concurrently. It records run history in
SQLite and combined output in log files under `/var/lib/server-maintenance`,
and supports cancellation of active task process groups.
The image contains application code only. Mount these paths at runtime:
@@ -77,6 +77,7 @@ readable. `description` is optional on a task and otherwise comes from the
first non-empty line of its module docstring. The order of keys in `tasks`
controls the listed tasks' display order. Include a task as `{}` when only
ordering is needed. Unlisted direct Python files follow alphabetically.
Use `\n` for a line break or `\n\n` for a paragraph gap in descriptions.
Each task declaration may contain `title`, `description`, `inputs`,
`wait_for_result`, `download_artifacts`, and `execution`. Set
@@ -143,9 +144,14 @@ Tasks can write files to the directory named by
`SERVER_MAINTENANCE_ARTIFACTS`. Run-detail pages show them when the task
finishes. JSON artifacts offer an **Open** link that renders the JSON in a new
tab and leaves the artifact available; other artifacts offer a download link.
The server deletes an artifact after its download transfer ends; any artifact
not downloaded is removed after 24 hours. A browser cannot confirm that a
downloaded file was retained on the client device.
The server deletes an artifact after its download stream completes; interrupted
downloads remain available for retry. Active runs' artifacts never expire.
After a run finishes, artifacts expire after 24 hours. Failed or cancelled
runs keep valid JSON diagnostic reports, but discard incomplete downloads.
Ordinary input files are removed when execution ends; saved inputs remain in
SQLite for history and review replay. Run pages fetch only appended log bytes.
A browser cannot confirm that a downloaded file was retained on the client device.
## Background tasks
+43 -17
View File
@@ -10,6 +10,7 @@ import subprocess
import threading
import time
from collections.abc import Mapping
from datetime import datetime
from pathlib import Path
from .config import Settings
@@ -127,31 +128,54 @@ class RunManager:
runs_root = self.settings.state_root / "runs"
cutoff = time.time() - artifact_max_age_seconds
if runs_root.is_dir():
lifetimes = self.store.run_lifetimes()
for path in runs_root.iterdir():
if not path.is_dir():
continue
run_id_text, separator, kind = path.name.partition(".")
if not separator or not run_id_text.isdigit() or kind not in {"uploads", "artifacts"}:
if not separator or kind not in {"uploads", "artifacts", "input.json", "execution-input.json"} or not run_id_text.isdigit() or path.is_symlink():
continue
run = self.store.get(int(run_id_text))
if kind == "uploads" and (run is None or run.status not in {"queued", "running"}):
run_id = int(run_id_text)
if run_id not in lifetimes:
# A worker may have created this run after the bulk snapshot.
run = self.store.get(run_id)
lifetimes[run_id] = (run.status, run.finished_at) if run else (None, None)
status, finished_at = lifetimes[run_id]
if status in {"queued", "running"}:
continue
if kind in {"input.json", "execution-input.json"} and path.is_file():
path.unlink(missing_ok=True)
elif kind == "uploads" and path.is_dir():
shutil.rmtree(path, ignore_errors=True)
elif kind == "artifacts":
elif kind == "artifacts" and path.is_dir():
if status in {"failed", "interrupted", "cancelled"}:
self._keep_diagnostic_reports(path)
try:
expired = path.stat().st_mtime < cutoff
completed_at = datetime.fromisoformat(finished_at).timestamp() if finished_at else path.stat().st_mtime
except FileNotFoundError:
continue
if expired:
if completed_at < cutoff:
shutil.rmtree(path, ignore_errors=True)
for path in runs_root.glob("*.execution-input.json"):
run_id_text = path.name.removesuffix(".execution-input.json")
if not run_id_text.isdigit():
continue
run = self.store.get(int(run_id_text))
if run is None or run.status not in {"queued", "running"}:
path.unlink(missing_ok=True)
self._cleanup_orphaned_task_state(orphaned_task_state_retention_seconds)
@staticmethod
def _keep_diagnostic_reports(root: Path) -> None:
try:
paths = list(root.iterdir())
except FileNotFoundError:
return
for path in paths:
if path.is_file() and not path.is_symlink() and path.suffix.casefold() == ".json":
try:
json.loads(path.read_text(encoding="utf-8"))
continue
except FileNotFoundError:
continue
except ValueError:
pass
if path.is_dir() and not path.is_symlink():
shutil.rmtree(path)
else:
path.unlink(missing_ok=True)
def _cleanup_orphaned_task_state(self, retention_seconds: int) -> None:
"""Delete state only after a group remains absent for a grace period."""
@@ -318,6 +342,7 @@ class RunManager:
environment = self._environment(task.group_id)
environment["SERVER_MAINTENANCE_INPUT"] = str(input_path)
environment["SERVER_MAINTENANCE_ARTIFACTS"] = str(artifact_root)
self.store.set_log_path(run.id, str(log_path))
try:
with log_path.open("wb") as log_file:
@@ -347,11 +372,12 @@ class RunManager:
self._processes.pop(run.id, None)
self._delete_uploads(run.id)
self._delete_execution_input(run.id)
(self.settings.state_root / "runs" / f"{run.id}.input.json").unlink(missing_ok=True)
if self.store.is_cancel_requested(run.id) or (
parent_run_id is not None and self.store.is_cancel_requested(parent_run_id)
):
shutil.rmtree(artifact_root, ignore_errors=True)
self._keep_diagnostic_reports(artifact_root)
return self.store.finish(
run.id,
status="cancelled",
@@ -365,7 +391,7 @@ class RunManager:
log_path=str(log_path),
)
if completed.status != "succeeded":
shutil.rmtree(artifact_root, ignore_errors=True)
self._keep_diagnostic_reports(artifact_root)
else:
try:
artifact_root.rmdir()
+11
View File
@@ -256,6 +256,17 @@ class RunStore:
with self._connect() as owned_connection:
return self.get(run_id, connection=owned_connection)
def run_lifetimes(self) -> dict[int, tuple[str, str | None]]:
with self._connect() as connection:
return {
row["id"]: (row["status"], row["finished_at"])
for row in connection.execute("SELECT id, status, finished_at FROM runs")
}
def set_log_path(self, run_id: int, path: str) -> None:
with self._connect() as connection:
connection.execute("UPDATE runs SET log_path = ? WHERE id = ?", (path, run_id))
def last_run(self, target_kind: str, target_path: str) -> Run | None:
with self._connect() as connection:
row = connection.execute(
+33 -9
View File
@@ -2,6 +2,7 @@
from __future__ import annotations
import codecs
import ipaddress
import json
from datetime import datetime
@@ -137,11 +138,13 @@ def create_app(settings: Settings | None = None) -> Flask:
if run is None:
abort(404)
execution_task = execution_replay_task(run)
output, offset, _, _ = read_run_output(run)
return render_template(
"run.html",
run=run,
children=store.children(run.id),
output=read_run_output(run),
output=output,
output_offset=offset,
execution_task=execution_task,
)
@@ -195,15 +198,36 @@ def create_app(settings: Settings | None = None) -> Flask:
run = store.get(run_id)
if run is None:
abort(404)
return {"output": read_run_output(run)}
raw_offset = request.args.get("offset")
if raw_offset is None:
return {"output": read_run_output(run)[0]}
if not raw_offset.isdecimal() or len(raw_offset) > 19 or int(raw_offset) > 2**63 - 1:
abort(400)
output, offset, reset, has_more = read_run_output(run, int(raw_offset), 1024 * 1024)
return {
"output": output, "offset": offset, "reset": reset, "has_more": has_more,
"status": run.status,
}
def read_run_output(run) -> str:
def read_run_output(run, offset: int = 0, maximum_bytes: int | None = None) -> tuple[str, int, bool, bool]:
if not run.log_path:
return ""
return "", 0, offset != 0, False
try:
return Path(run.log_path).read_text(encoding="utf-8", errors="replace")
with Path(run.log_path).open("rb") as stream:
stream.seek(0, 2)
size = stream.tell()
reset = offset > size
if reset:
offset = 0
stream.seek(offset)
data = stream.read() if maximum_bytes is None else stream.read(maximum_bytes)
except FileNotFoundError:
return ""
return "", 0, offset != 0, False
has_more = offset + len(data) < size
decoder = codecs.getincrementaldecoder("utf-8")("replace")
output = decoder.decode(data, final=run.status not in {"queued", "running"} and not has_more)
pending, _ = decoder.getstate()
return output, offset + len(data) - len(pending), reset, has_more
@app.get("/runs/<int:run_id>/artifacts")
def run_artifacts(run_id: int):
@@ -224,7 +248,7 @@ def create_app(settings: Settings | None = None) -> Flask:
@app.get("/runs/<int:run_id>/artifacts/<filename>/view")
def view_artifact(run_id: int, filename: str):
run = store.get(run_id)
if run is None or run.status != "succeeded" or not filename.casefold().endswith(".json"):
if run is None or run.status in {"queued", "running"} or not filename.casefold().endswith(".json"):
abort(404)
artifact = manager.artifact_path(run.id, filename)
if artifact is None:
@@ -234,7 +258,7 @@ def create_app(settings: Settings | None = None) -> Flask:
@app.get("/runs/<int:run_id>/artifacts/<filename>")
def download_artifact(run_id: int, filename: str):
run = store.get(run_id)
if run is None or run.status != "succeeded":
if run is None or run.status in {"queued", "running"} or (run.status != "succeeded" and not filename.casefold().endswith(".json")):
abort(404)
artifact = manager.artifact_path(run.id, filename)
if artifact is None:
@@ -250,11 +274,11 @@ def create_app(settings: Settings | None = None) -> Flask:
def delete_after_transfer():
try:
yield from original_response
manager.delete_artifact(run.id, artifact.name)
finally:
close = getattr(original_response, "close", None)
if close is not None:
close()
manager.delete_artifact(run.id, artifact.name)
response.response = delete_after_transfer()
response.direct_passthrough = False
+33 -14
View File
@@ -1,25 +1,33 @@
const output = document.querySelector("#run-output");
if (output) {
let refreshTimer;
let offset = Number(output.dataset.outputOffset || 0);
let detailsFinalized = !["queued", "running"].includes(
document.querySelector("#run-status")?.textContent.trim(),
);
const refresh = async () => {
const response = await fetch(output.dataset.outputUrl, { cache: "no-store" });
if (response.ok) {
const nextOutput = (await response.json()).output;
if (nextOutput !== output.textContent) {
output.textContent = nextOutput;
let delay = 2000;
try {
const url = new URL(output.dataset.outputUrl, window.location.href);
url.searchParams.set("offset", String(offset));
const response = await fetch(url, { cache: "no-store" });
if (!response.ok) {
return;
}
const state = await response.json();
if (state.reset) {
output.replaceChildren();
}
if (state.output) {
output.append(document.createTextNode(state.output));
}
offset = state.offset;
const status = document.querySelector("#run-status");
if (status) {
const stateResponse = await fetch(status.dataset.stateUrl, { cache: "no-store" });
if (stateResponse.ok) {
const state = await stateResponse.json();
status.textContent = state.status;
if (state.status !== "queued" && state.status !== "running") {
if (state.has_more) {
delay = 0;
} else if (state.status !== "queued" && state.status !== "running") {
if (!detailsFinalized) {
const pageResponse = await fetch(window.location.href, { cache: "no-store" });
if (!pageResponse.ok) {
@@ -36,13 +44,22 @@ if (output) {
nextDetails.after(...page.querySelectorAll(actions));
detailsFinalized = true;
}
window.clearInterval(refreshTimer);
const artifacts = document.querySelector("#run-artifacts");
if (artifacts) {
await refreshArtifacts(artifacts);
}
delay = null;
}
}
} catch (error) {
console.warn("Could not refresh run details:", error);
} finally {
if (delay !== null) {
window.setTimeout(refresh, delay);
}
}
};
refresh();
refreshTimer = window.setInterval(refresh, 2000);
}
const refreshArtifacts = async (container) => {
@@ -73,8 +90,9 @@ const refreshArtifacts = async (container) => {
};
for (const container of document.querySelectorAll("#run-artifacts")) {
if (!output) {
refreshArtifacts(container);
window.setInterval(() => refreshArtifacts(container), 2000);
}
}
for (const choice of document.querySelectorAll("button[data-execution-choice]")) {
@@ -154,6 +172,7 @@ for (const form of document.querySelectorAll("form[data-wait-for-result]")) {
};
const waitForRun = async (run) => {
const started = Date.now();
while (true) {
const response = await fetch(run.state_url, { cache: "no-store" });
if (!response.ok) {
@@ -171,7 +190,7 @@ for (const form of document.querySelectorAll("form[data-wait-for-result]")) {
}
return;
}
await new Promise((resolve) => window.setTimeout(resolve, 500));
await new Promise((resolve) => window.setTimeout(resolve, Date.now() - started < 10000 ? 500 : 2000));
}
};
+1 -1
View File
@@ -37,5 +37,5 @@
<ul></ul>
</section>
<h2>Output</h2>
<pre id="run-output" data-output-url="{{ url_for('run_output', run_id=run.id) }}">{{ output }}</pre>
<pre id="run-output" data-output-url="{{ url_for('run_output', run_id=run.id) }}" data-output-offset="{{ output_offset }}">{{ output }}</pre>
{% endblock %}
@@ -0,0 +1,87 @@
import pytest
from app.web import create_app
from .conftest import write_task
from .test_runner import wait_for, wait_for_output
@pytest.fixture
def output_run(settings):
app = create_app(settings)
store = app.extensions["run_store"]
run, _ = store.enqueue(target_kind="task", target_path="example/task.py", queue_group="example")
store.claim_next("example")
path = settings.state_root / "test.log"
path.write_bytes(b"first\n")
with store._connect() as connection:
connection.execute("UPDATE runs SET log_path = ? WHERE id = ?", (str(path), run.id))
return app.test_client(), run.id, path, store
def test_log_polling_returns_only_appended_bytes_and_status(output_run):
client, run_id, path, store = output_run
first = client.get(f"/runs/{run_id}/output?offset=0").get_json()
assert first == {"output": "first\n", "offset": 6, "reset": False, "has_more": False, "status": "running"}
with path.open("ab") as stream:
stream.write("next: \u20ac\n".encode())
store.finish(run_id, status="succeeded")
next_chunk = client.get(f"/runs/{run_id}/output?offset=6").get_json()
assert next_chunk["output"] == "next: \u20ac\n"
assert next_chunk["offset"] == path.stat().st_size
assert next_chunk["status"] == "succeeded"
assert client.get(f"/runs/{run_id}/output").get_json() == {"output": "first\nnext: \u20ac\n"}
def test_incomplete_utf8_is_deferred_until_the_next_poll(output_run):
client, run_id, path, _ = output_run
path.write_bytes(b"ok\xe2\x82")
first = client.get(f"/runs/{run_id}/output?offset=0").get_json()
assert first["output"] == "ok"
assert first["offset"] == 2
with path.open("ab") as stream:
stream.write(b"\xac")
second = client.get(f"/runs/{run_id}/output?offset=2").get_json()
assert second["output"] == "\u20ac"
assert second["offset"] == 5
def test_truncated_log_resets_the_browser_offset(output_run):
client, run_id, path, _ = output_run
path.write_bytes(b"new")
result = client.get(f"/runs/{run_id}/output?offset=6").get_json()
assert result["reset"] is True
assert result["output"] == "new"
assert result["offset"] == 3
def test_large_log_is_drained_in_bounded_chunks(output_run):
client, run_id, path, store = output_run
path.write_bytes(b"x" * (1024 * 1024 - 1) + "\u20ac".encode())
store.finish(run_id, status="succeeded")
first = client.get(f"/runs/{run_id}/output?offset=0").get_json()
assert first["has_more"] is True
assert first["offset"] == 1024 * 1024 - 1
second = client.get(f"/runs/{run_id}/output?offset={first['offset']}").get_json()
assert second["output"] == "\u20ac"
assert second["has_more"] is False
@pytest.mark.parametrize("offset", ["-1", "abc", "1.5", "9" * 100, str(2**63)])
def test_invalid_log_offset_is_rejected(output_run, offset):
client, run_id, _, _ = output_run
assert client.get(f"/runs/{run_id}/output?offset={offset}").status_code == 400
def test_running_task_output_is_visible_before_completion(settings):
write_task(settings.task_root, "example", "long.py", 'import time\nprint("working", flush=True)\ntime.sleep(30)\n')
app = create_app(settings)
manager = app.extensions["run_manager"]
run, _ = manager.run_task(app.extensions["catalog"].task("example/long.py"))
try:
wait_for_output(settings.state_root / "runs" / f"{run.id}.log", "working")
response = app.test_client().get(f"/runs/{run.id}/output?offset=0").get_json()
assert response["status"] == "running"
assert response["output"] == "working\n"
finally:
manager.request_cancel(run.id)
wait_for(manager, run.id, {"cancelled"})
@@ -0,0 +1,95 @@
import json
import os
import time
from datetime import UTC, datetime, timedelta
from app.web import create_app
from .conftest import write_task
from .test_runner import wait_for
def test_active_inputs_and_artifacts_never_expire(manager, monkeypatch):
run, _ = manager.store.enqueue(target_kind="task", target_path="example/task.py", queue_group="example")
manager.store.claim_next("example")
root = manager.settings.state_root / "runs"
for name in (f"{run.id}.artifacts", f"{run.id}.uploads"):
path = root / name
path.mkdir(parents=True)
(path / "data").write_text("active")
os.utime(path, (time.time() - 3 * 86400,) * 2)
input_path = root / f"{run.id}.input.json"
input_path.write_text("{}")
monkeypatch.setattr(manager.store, "run_lifetimes", lambda: {})
manager.cleanup_transient_files()
assert input_path.exists()
assert (root / f"{run.id}.artifacts/data").exists()
assert (root / f"{run.id}.uploads/data").exists()
def test_artifact_retention_starts_at_completion_not_directory_creation(manager):
run, _ = manager.store.enqueue(target_kind="task", target_path="example/task.py", queue_group="example")
manager.store.finish(run.id, status="succeeded")
root = manager.settings.state_root / "runs"
artifact = root / f"{run.id}.artifacts"
artifact.mkdir(parents=True)
(artifact / "report.json").write_text("{}")
os.utime(artifact, (time.time() - 3 * 86400,) * 2)
(root / f"{run.id}.input.json").write_text("{}")
manager.cleanup_transient_files()
assert (artifact / "report.json").exists()
assert not (root / f"{run.id}.input.json").exists()
old = (datetime.now(UTC) - timedelta(days=3)).isoformat()
with manager.store._connect() as connection:
connection.execute("UPDATE runs SET finished_at = ? WHERE id = ?", (old, run.id))
manager.cleanup_transient_files()
assert not artifact.exists()
def test_failed_task_keeps_valid_diagnostics_but_not_incomplete_downloads(manager):
write_task(manager.settings.task_root, "example", "fail.py", '''
import os
from pathlib import Path
root = Path(os.environ["SERVER_MAINTENANCE_ARTIFACTS"])
(root / "report.json").write_text('{"completed_steps": 2}')
(root / "broken.json").write_text('{')
(root / "partial.zip").write_bytes(b"partial")
raise SystemExit(1)
''')
task = manager.catalog.task("example/fail.py")
run, _ = manager.run_task(task)
wait_for(manager, run.id, {"failed"})
assert manager.artifacts(run.id) == ["report.json"]
assert not (manager.settings.state_root / "runs" / f"{run.id}.input.json").exists()
def test_failed_json_report_is_accessible(settings):
app = create_app(settings)
manager = app.extensions["run_manager"]
run, _ = manager.store.enqueue(target_kind="task", target_path="example/fail.py", queue_group="example")
manager.store.finish(run.id, status="failed")
root = settings.state_root / "runs" / f"{run.id}.artifacts"
root.mkdir(parents=True)
(root / "report.json").write_text('{"completed_steps":2}')
client = app.test_client()
response = client.get(f"/runs/{run.id}/artifacts/report.json/view")
assert response.status_code == 200
assert json.loads(response.data) == {"completed_steps": 2}
def test_aborted_artifact_download_can_be_retried(settings):
app = create_app(settings)
manager = app.extensions["run_manager"]
run, _ = manager.store.enqueue(target_kind="task", target_path="example/archive.py", queue_group="example")
manager.store.finish(run.id, status="succeeded")
root = settings.state_root / "runs" / f"{run.id}.artifacts"
root.mkdir(parents=True)
path = root / "review.zip"
path.write_bytes(b"x" * 100_000)
client = app.test_client()
response = client.get(f"/runs/{run.id}/artifacts/review.zip", buffered=False)
next(iter(response.response))
response.close()
assert path.exists()
retry = client.get(f"/runs/{run.id}/artifacts/review.zip")
assert len(retry.data) == 100_000
assert not path.exists()