forge_ops_tracker 0.11.0 copy "forge_ops_tracker: ^0.11.0" to clipboard
forge_ops_tracker: ^0.11.0 copied to clipboard

Published in the last hour
SDKDartFlutter
PlatformAndroidiOSLinuxmacOSWindows
→

Metadata

Dart error reporting client for ForgeOps. Works in any Dart VM environment (server apps, CLI tools, Flutter mobile/desktop): see README.md for why Flutter Web specifically isn't covered, and for the s [...]

More...

Changelog #

0.11.0 - 2026-09-25 #

  • A database span can now carry the SQL it ran, such as a local SQLite query: span, spanAsync and recordSpan take optional statement and dbSystem (such as 'sqlite') arguments. The statement is masked before it leaves the process (every string and number literal becomes ?), cut at 4000 characters, and sent in the span's data as db.statement, with db.system lowercased. A db.statement put in data directly is masked the same way. Both are ignored on spans of any other kind.

0.10.0 - 2026-09-25 #

  • Change tracking: recordChange(kind, title, {details, environment, service, actor, url, id, occurredAt}) records one change you made (a feature flag flipped, a config value changed, a migration run) so ForgeOps can show it next to the errors and slowdowns that followed. kind is one of feature_flag, config, migration, dependency, infrastructure or other (exported as changeKinds); anything else is sent as other. The title is cut to 200 characters, environment defaults to the configured one, and occurredAt to now. Queued on the event loop like an error report, so it never blocks and never throws; a failed delivery, including a 403 on a plan without change tracking, is dropped quietly. await flushChanges() before a short-lived program exits. This client sends no startup snapshot of its own.

0.9.0 #

  • Trace context: an error captured inside a trace/traceAsync (across awaits) now carries the trace's trace_id, and so does an error that escaped a trace and is captured afterward, which links it to errors other services reported for the same request (the two projects must be linked in ForgeOps). New httpSpan/httpSpanAsync(method, url, (headers) => ...) record an outgoing call as an http span and hand your code a W3C traceparent header naming that span, so the service being called continues the trace. trace/traceAsync take an optional traceparent to continue a trace another service started (strictly parsed; a bad value starts a fresh trace), and currentTraceId() returns the open trace's id. Trace and span ids are W3C ids, never all zeros; TraceParent is exported. Two new options: propagateTraces (default true; false stops the header but still records the span) and tracePropagationTargets (default null, every host; otherwise host strings matched exactly or as a subdomain on a dot boundary, and RegExps matched against the host). With trackTracing off a trace still has an id for errors and headers; only its spans stop being sent. Nothing instruments HttpClient automatically, and errors captured outside a trace are unchanged.

0.8.0 #

  • Database errors now say where to look. When an error carries the SQL behind a failed local database call (package:sqlite3's SqliteException.causingStatement, or sqflite's and SQLite's error text), the event carries the names of the tables and views (and any stored procedure) that SQL touched. On by default (captureSqlObjects); names are identifiers, never values. New opt-in captureSqlStatement (default false) also sends the statement itself, with every string and number replaced by ?. Each project has its own server-side setting that can stop the statement being stored regardless of this flag; the names are still kept.

0.7.1 #

  • Documentation only: the README and package description now describe ForgeOps as a hosted service, link to getforgeops.net, and show the real host in the connection string example instead of a placeholder. No code changes.

0.7.0 #

  • Performance percentiles: every performance sample now carries a small latency histogram alongside its count/sum/max (fixed buckets of 50, 100, 250, 500, 1000, 2500, 5000 and 10000ms, plus an overflow bucket), so ForgeOps can show an approximate p50/p95/p99 per transaction instead of only an average. No new config; this rides the existing performance tracking flag and flush interval.

0.6.0 #

  • Custom metrics and infrastructure monitoring: captureMetric(name, [value = 1]) records a named business event (a purchase, a level completed) and captureInfrastructureMetric(name, value, {hostname}) a CPU/memory/disk reading from one of your own hosts. Buffered and flushed as one batch once metricFlushInterval/infrastructureMetricFlushInterval has elapsed at a capture (no background timer, for the same reason performance monitoring has none); await flushMetrics() before a short-lived program exits. Nothing is automatic, so there is no trackMetrics flag.

0.5.0 #

  • Distributed tracing: trace/traceAsync wrap a unit of work, span/spanAsync add child spans (nested via Zone, so parallel futures each parent under their own span), and recordSpan adds one you timed yourself. A trace is sent to the new /spans endpoint only when it took at least Configuration.traceCaptureThreshold (one second), and is sent even if the body throws. Nothing is traced automatically (no web framework integration in this client). Delivery is the same bounded event-loop queue as error reports; await flushSpans() before a short-lived program exits. Turn it off with Configuration.trackTracing = false.

0.4.0 #

  • Performance monitoring: recordPerformance tallies how long named work took (count, total, max) and flushes one small aggregate per transaction, for the Performance page's per-transaction table. timeTransaction/timeTransactionAsync wrap a block and record even if it throws. Nothing is timed automatically (no web framework integration in this client). There is deliberately no background timer, since a pending Timer would stop a Dart CLI program from exiting on its own: a flush starts from a recordPerformance call once performanceFlushInterval has elapsed, or from an explicit flushPerformance(), which a short-lived program or a Flutter app's AppLifecycleState.paused handler should call. Turn it off with Configuration.trackPerformance = false.

0.3.0 #

  • Breadcrumbs: addBreadcrumb records a small, bounded trail of recent events (30 by default, configurable via Configuration.maxBreadcrumbs) attached to whatever error gets reported next. addBreadcrumb is a plain global by default (the same single-flow reasoning setUser already documents); runWithBreadcrumbs scopes a fresh, isolated trail to one call chain via Zone instead, for a Dart server handling more than one request at a time on the same isolate. Never recorded automatically (no web framework built into this client to record one from); turn it off entirely with Configuration.trackBreadcrumbs = false.

0.2.0 #

  • Identify the affected user on a reported error: pass user to captureException, or call setUser/runWithUser to attach one to every subsequently reported error until changed or cleared. setUser is a plain global (correct for a single-user Flutter/CLI app); runWithUser scopes the value to one call chain via Dart's Zone mechanism instead, for a Dart server handling more than one request at a time on the same isolate. Never scrubbed by PII scrubbing.

0.1.0 #

  • Initial release: works in any Dart VM environment (server apps, CLI tools, Flutter mobile/desktop) via dart:io/dart:convert, zero runtime dependencies. PII scrubbing (email addresses, formatted SSNs/credit cards, known API key/token formats, and any field whose name suggests a secret) runs before a payload ever leaves the host app's process. Delivery runs on an async drain loop with a bounded queue and short HTTP timeouts; every failure mode is caught and dropped rather than thrown, so a broken or unreachable tracker can never take down the host app. See the separate flutter_forge_ops_tracker package for Flutter-specific automatic capture (FlutterError.onError/PlatformDispatcher.onError) on top of this one.

← Metadata

0
likes
130
points
282
downloads

Documentation

API reference

Publisher

unverified uploader

Weekly Downloads

Metadata

Dart error reporting client for ForgeOps. Works in any Dart VM environment (server apps, CLI tools, Flutter mobile/desktop): see README.md for why Flutter Web specifically isn't covered, and for the separate flutter_forge_ops_tracker package that adds Flutter-specific automatic capture on top of this one.

Homepage

Topics

#error-reporting #logging #monitoring #observability

License

MIT (license)

More

Packages that depend on forge_ops_tracker

Back