Commit graph

1,499 commits

Author SHA1 Message Date
Alex Garcia
d7b7d7571b
Add a plugin telemetry kit: public registry API, linked_root_span_kwargs, test helpers, docs
A survey of five plugin OTel plans (datasette-paper, -agent, -litestream,
-accounts, -cron) found every one hand-copying the same core machinery:
the registry classes, the conformance-test harness, the pytest fixtures,
the bucket boundaries and the detached-root-with-Link recipe. This makes
that machinery importable instead:

- The registry classes are documented public API. Attribute gains
  values= (a closed enum the conformance helpers enforce - what makes an
  attribute safe as a metric dimension); SpanName gains prefix=True for
  span families like "chat {model}" whose names share a fixed prefix,
  matched by span_for() after exact names. span_for()/attribute helpers
  accept a spans= tuple so plugin registries can use them.
- datasette.telemetry.linked_root_span_kwargs(): the root-span-with-Link
  shape for work a request caused without containing - background jobs,
  scheduled ticks, block=False writes. Core's own write thread now uses
  it instead of building the kwargs inline.
- datasette.telemetry_testing: the session provider fixtures, otel_spans
  / otel_metrics, a two-way registry conformance checker (including enum
  and prefix handling, filtered by instrumentation scope) and an
  assert_package_never_imports_sdk() guard. Core's conftest now imports
  these instead of defining them, so the suite consumes the kit exactly
  as a plugin's would.
- New "Telemetry for plugin authors" docs page: scope discipline,
  registry usage, privacy/cardinality rules, named-callable guidance,
  request_span(), the background root-with-link convention (one root per
  tick, always emitted), provider-ordering facts and known caveats.
  request_span() is now documented public API.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012U7coQfVu8nK2R4q2mCULA
2026-09-24 09:34:14 -07:00
Alex Garcia
2cefab7b4c
Count callback-style calls in db.client.operation.duration
The callback entry points gained db.query spans in the database-spans PR;
this adds their other half - the duration histogram measurement, so a
plugin's execute_fn/execute_write_fn work and the JSON write API's inserts
and deletes stop being invisible to the one series that survives trace
sampling. execute_isolated_fn records "write" when the database is mutable
(the call blocks the write queue) and "read" when immutable (it runs on
the read pool). error.type comes from the raised exception class, same as
the SQL-string paths.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012U7coQfVu8nK2R4q2mCULA
2026-09-24 09:14:06 -07:00
Alex Garcia
8f49def8cb
Review polish: fix stale exemplar context, dedupe rationales, close test gaps
- The exemplars docs described "the pinned opentelemetry-exporter-prometheus"
  and "Datasette's own Prometheus exporter" - context from demo/plugin work
  that is no longer part of this stack. Reworded to stand alone.
- Saturate a num_sql_threads=1 pool and assert the queue-depth gauge reads
  above zero - the headline alerting metric previously only had an absence
  test, and this also pins the private ThreadPoolExecutor._work_queue
  attribute it depends on.
- Pin error.type on the write path of db.client.operation.duration - the
  write wrappers time a different code path than the read one already tested.
- Isolate the non-threaded-mode gauge test from other live instances instead
  of comparing global observation counts, which a GC pass could shift.
- Halve the metrics banner, point conftest's meter note at it, compact the
  interrupted-counter call-site comment to a registry pointer, note why
  instrument and registry descriptions are separate strings, and stop
  calling the metric dimension a "later phase" now that metrics shipped.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012U7coQfVu8nK2R4q2mCULA
2026-09-24 09:14:06 -07:00
Alex Garcia
3024d950d4
Check metric attributes in the registry conformance test
Span attributes were checked in both directions; metric attributes were not
checked at all, so the generated reference could publish an incomplete list
with nothing to catch it.

The metric workload lives in an `emitted_metrics` fixture, mirroring the
span side, and error.type is checked like every other attribute rather than
exempted for being optional - the workload reaches it two separate ways.

(Adapted from b30c5341: the old workload's facet-timeout probe belongs to
phase 5 and is dropped, and the interrupted counter now needs a query that
exceeds the *configured* time limit - custom short budgets are excluded from
the count on this lineage - so the fixture runs one against a second
instance configured with sql_time_limit_ms=5.)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F2h9ANGZ7paWSpqs5DUAcG
2026-09-24 09:14:06 -07:00
Alex Garcia
383fa1ad0a
Register metrics and give histograms bucket boundaries suited to seconds
Both histograms declared unit="s" but inherited OpenTelemetry's default
boundaries, which are tuned for milliseconds - so every SQLite query
landed in the single (0, 5] second bucket and every quantile query
returned noise.

The boundaries are the semantic conventions' recommended set for
db.client.operation.duration plus 0.0001 and 0.0005 at the bottom, since
SQLite is in-process and many real queries take tens of microseconds.

(Adapted from 024f2029: that commit assumed the metrics were already in
telemetry_registry.py, which on this lineage held spans only - so this
commit also brings the MetricName registry machinery, the registry
entries for all eight phase-3 metrics, the cog-generated Metric
reference in internals.rst, and the datasette.operation attribute. The
template and facet histograms it also touched belong to phase 5 and are
not included.)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F2h9ANGZ7paWSpqs5DUAcG
2026-09-24 09:14:06 -07:00
Alex Garcia
90b727db73
Add OpenTelemetry metrics for SQL thread pool saturation and query latency
Spans describe requests that have finished. They structurally cannot answer
"am I saturating my 3 SQL threads right now", because that is a level rather
than an event - and with num_sql_threads defaulting to 3, it is usually the
first thing worth knowing about a busy Datasette. This adds the metrics that
answer it.

Five observable gauges, computed only when something is collecting, so an
instance with no MeterProvider installed does no work for them at all:

  datasette.sql.threads.limit         num_sql_threads
  datasette.sql.threads.queue_depth   queries waiting for a free thread
  datasette.sql.queries.pending       in-flight reads, by db.namespace
  datasette.write.queue_depth         writes behind the single write thread
  datasette.connections.open          tracked file connections

Three instruments recorded inline, which matters because metrics survive
trace sampling and spans do not - an operator sampling 1% of traces still
gets 100% of the latency distribution:

  db.client.operation.duration        semconv histogram, with error.type
  datasette.write.queue_wait          the metric twin of the existing span
  datasette.sql.queries.interrupted   sql_time_limit_ms kills

The interrupted counter closes a gap the plan called out as unanswerable:
"how often are we killing queries at the limit" is a rate, and a rate cannot
be recovered from sampled spans.

Core still creates no provider of any kind, so the architecture is unchanged;
`grep -rn 'opentelemetry.sdk' datasette/` stays empty. One real difference
from tracing is worth recording: _ProxyMeter and its instruments forward to a
provider installed after they were created, whereas ProxyTracer permanently
caches the first concrete tracer it resolves. Module-level instruments are
therefore safe and the test fixture has no ordering constraint.

Live instances are tracked in a lock-guarded WeakSet so instrumenting an
instance never keeps it alive. The pool gauges carry no attribute saying
which Datasette produced them: production runs one instance per process, and
adding an id to disambiguate the test suite's hundreds of instances would buy
unbounded attribute cardinality to fix a case that does not occur. The
collision is documented instead, and the gauge callbacks are plain generator
functions so tests can assert exact values by calling them directly rather
than through the SDK's last-value aggregation.

demos/otel/metrics_demo.py fires 12 concurrent 40ms queries at a 3-thread
pool and samples the gauges mid-flight: queue_depth peaks at exactly 9, and
the duration histogram reads max=0.1695s for a query whose work is 40ms. That
gap is the queue, and it is the thing traces alone will not show you.

Also corrects the demo README's privacy section, which still claimed
parameter values are never recorded - that stopped being unconditionally true
when trace_sql_parameters landed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

(cherry picked from 6ef0dd8c and adapted to the rebuilt phase-1 stack:
attribute names now come from telemetry_registry where entries exist, the
meter carries the instrumentation-scope version and schema URL, and the
interrupted-queries counter skips expected timeouts - callers that opted
into a deliberately short budget, like facet suggestion - matching how
those are excluded from span error status. The internals.rst reference
lands with the registry commit that follows.)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F2h9ANGZ7paWSpqs5DUAcG
2026-09-24 09:14:06 -07:00
Alex Garcia
9bce2eec7c
Emit an OpenTelemetry span for each HTTP request
PR #2863
2026-09-24 09:14:03 -07:00
Alex Garcia
daf1989ebe
OpenTelemetry tracing for the database layer (#2862)
PR #2862
2026-09-24 09:11:08 -07:00
Simon Willison
3e9356dde0 Rename modal trigger option to returnFocusTo, refs #2790 2026-09-17 15:46:34 -07:00
Simon Willison
ad1d0bf6f9 Rename modal dismissal reason to source, refs #2790 2026-09-17 15:32:47 -07:00
Simon Willison
71600f1c0a Simplify shared modals now that all dialogs use the document, refs #2790 2026-09-17 14:10:45 -07:00
Simon Willison
15d511e2da Render the column chooser without shadow DOM, refs #2790 2026-09-17 14:10:45 -07:00
Simon Willison
e60d1bfe1c Render navigation search without shadow DOM, refs #2790 2026-09-17 14:10:44 -07:00
Simon Willison
3b013b7ea3 Refactor row editing and insertion to use the shared modal, refs #2790 2026-09-17 13:28:26 -07:00
Simon Willison
90f543327e Refactor row deletion to use the shared modal, refs #2790 2026-09-17 13:28:26 -07:00
Simon Willison
814165c8b1 Refactor the alter table dialog to use the shared modal, refs #2790 2026-09-17 13:28:25 -07:00
Simon Willison
c82a98c88a Refactor the create table dialog to use the shared modal, refs #2790 2026-09-17 13:28:25 -07:00
Simon Willison
de37f1451f Refactor the column type dialog to use the shared modal, refs #2790 2026-09-17 13:28:24 -07:00
Simon Willison
17b19b4d27 Refactor mobile column actions to use the shared modal, refs #2790 2026-09-17 13:28:24 -07:00
Simon Willison
328b2e6c6f Refactor the column chooser to use the shared modal, refs #2790 2026-09-17 13:28:24 -07:00
Simon Willison
c410ed9555 Refactor navigation search to use the shared modal, refs #2790 2026-09-17 13:28:23 -07:00
Simon Willison
e1f89494f1 Add reusable modal component, documentation and lifecycle tests, refs #2790 2026-09-17 13:28:23 -07:00
Simon Willison
3239d10b9b Use cog to show default values in settings.rst 2026-09-17 12:30:47 -07:00
Simon Willison
9cdf95ac2c Show correct query timings, closes #2446 2026-09-16 21:23:54 -07:00
Simon Willison
caf238aac8 Fix for GHSA-h547-rmjf-5m2m 2026-09-16 16:43:34 -07:00
Simon Willison
d2098e9f84
Fix facet selection for explicit exact filters
`colname__exact=value` is now treated the same as `colname=value` when determining which facets are selected.

Closes #1695
2026-09-16 14:56:09 -07:00
Simon Willison
266eaddb73
Fix remaining pytest warnings (#2928)
- Close in-memory database connections, including reads opened on threads
- Close completed and partial file upload file handles
- Close SQLite connections and file handles owned by tests
2026-09-16 14:50:06 -07:00
Simon Willison
0d0bb5dd8c
Merge branch 'main' into codex/test-write-wrapper-alter-rollback 2026-09-16 14:42:20 -07:00
Simon Willison
6c119323d0 Keep alter-table changes rollbackable by write wrappers 2026-09-16 14:40:50 -07:00
Simon Willison
6dd5297b34
Allow extra_template_vars to resolve to None
Closes #2005
2026-09-16 10:30:11 -07:00
Simon Willison
df4c0fde0f
Make request header lookups case-insensitive
Closes #1861
2026-09-16 10:27:35 -07:00
Simon Willison
faff4c8820 Return CSV errors as plain text, closes #2129 2026-09-15 15:53:10 -07:00
Rami Abdelrazzaq
82ed8d47f9
Fix float coercion for numeric filter parameters (#2876)
Closes #1681
2026-09-15 13:14:29 -07:00
AnkitaAdvitot
920adc6880
Fix infinite loop when streaming CSV on SQL views, fixes #2902 (#2903)
When streaming a view as CSV with ?_stream=on, _next was being
re-read from request.args, overwriting the kwarg passed by the
streaming loop with None. This caused the pagination offset to stay
stuck on page 1 forever.
2026-09-15 12:57:09 -07:00
Simon Willison
b09976b924 Tests for table count endpoint, refs #2914
Forgot to include these tests in commit 1538832830
2026-09-15 12:21:36 -07:00
Alex Garcia
374b194ff5
Add /-/tasks introspection endpoint for supervised background tasks (#2892)
Co-authored-by: Simon Willison <swillison@gmail.com>
2026-09-15 11:56:53 -07:00
Alex Garcia
784695aea6 Ensure startup() runs before any plugin ASGI middleware (#2891) 2026-09-15 11:22:45 -07:00
Alex Garcia
cca08d2886 Add shutdown() plugin hook with ordered graceful teardown (#2890) 2026-09-15 11:09:57 -07:00
Alex Garcia
0c31778580 Add datasette.add_background_task() with supervised launch after startup (#2889) 2026-09-15 10:55:54 -07:00
Simon Willison
b0407867cc Tweak display of count all button, refs #2914 2026-09-15 10:16:16 -07:00
Simon Willison
1538832830 /db/table/-/count? endpoint, fixed 'count all' button
Closes #2914
2026-09-15 10:03:17 -07:00
Simon Willison
b338c6f5f6
Migrate from httpx to httpx2, closes #2879
https://claude.ai/code/session_01Xdqoneq8ddvruVZETo6rFf
2026-09-10 19:44:49 -07:00
Simon Willison
92c7d4b608 Limit derived-table permissions to one source hop
Simplify the solution to 5de0c1724e - avoid contextvar.
2026-09-10 16:52:52 -07:00
Simon Willison
f70edbfa60 Filter incoming foreign-key relationships by view permission 2026-09-10 16:52:52 -07:00
Simon Willison
b97bb5f016 Reconcile write-timeout regression with the per-call limit 2026-09-10 16:52:26 -07:00
Simon Willison
e036907fc3 Reject structured row writes to virtual and shadow tables 2026-09-10 16:52:25 -07:00
Simon Willison
3f8d8417f6 Inherit source permissions for FTS vocabulary tables 2026-09-10 16:52:25 -07:00
Simon Willison
d334539a1e Deny SQLite statistics table access through a default hook 2026-09-10 16:52:25 -07:00
Simon Willison
506c4bb522 Match table permission identities using SQLite case semantics 2026-09-09 08:39:03 -07:00
Alex Garcia
c6ba7b3298 Refuse API token creation from restricted actors 2026-09-09 08:38:40 -07:00