The microsecond ladder's first edge was 100us, which sat ABOVE the mass of
every instrument using it. Measured on devnet: 99.3% of job_queued_us
samples, 92.5% of job_running_us and 90.4% of getobject_lookup_us fell in
that first bucket. histogram_quantile then interpolated inside bucket 0 and
returned `quantile / fraction_in_bucket_0 x first_edge` -- p75/p95/p99 of
job_queued_us read 75.52/95.66/99.69us against a prediction of
75.53/95.67/99.70. Three-decimal agreement: those panels were reporting
arithmetic on the bucket edge, not latency.
The fix was already half-written. kSubMillisecondBoundaries had been parked
in MetricsRegistry.cpp as [[maybe_unused]] with a comment noting exactly this
problem for nodestore reads. Its edges are now folded into kMicrosecondBuckets
rather than deleted, so the parked intent is carried forward: 1..1000us
resolution where the mass is, upper edges unchanged so multi-second stalls
stay measurable.
Also moves the GetObject count and charge ladders into HistogramBuckets.h, so
all five ladders have one owner and one set of invariant tests (29 now).
Adds check_bucket_parity.py, wired into the existing OTel naming workflow.
The C++ millisecond ladder and the collector's spanmetrics ladder are
specified to agree over their shared range; they were identical when shipped,
then the collector side alone was extended and nothing noticed for eleven
phases. The check asserts containment rather than equality, because jobs
outlive spans -- jobq_updatepaths averages ~60s, which no span approaches, so
demanding equality would force a ceiling that censors it. Verified it rejects
a missing collector edge, a bogus in-range edge, and a return to the 5s
ceiling.
ledger-data-sync's "Job Queue Wait p95 By Type" moves off the beast
jobq_*_q_milliseconds pair onto job_queued_us filtered by job_type. Those
beast metrics are ms-quantised at the source (Event rounds up to a whole
millisecond), so 94-100% of their samples sat in the first bucket and no
ladder change could fix them. Note the label values are camelCase
(job_type="ledgerData"), not the lowercase metric-name fragments.
Both histogram-fed alert thresholds re-validated and left unchanged, with the
measured basis recorded so neither gets tuned against the old artefact: only
0.0022% of job_queued_us samples exceed the 1s threshold, and every edge
bracketing the 1000ms ios_latency threshold survived the ladder change.
Docs: the rpc_size "known issue -- tracked separately" notes in the runbook
and 09-data-collection-reference are now resolved notes, the stale 10-edge
span_duration bucket list is corrected to the collector's real 20, and the
runbook gains a "Reading A Histogram Percentile" section covering both
saturation traps and the expected discontinuity after a ladder change.
beast::insight::Event documents itself as carrying "a millisecond time, or
other integral value", but both backends assumed the first case: the OTel
bridge declared every instrument with unit `ms` and StatsD tagged every
sample `|ms`. One Event does not measure time -- ServerHandler's "size"
records the serialized RPC response length -- so it exported as
rpc_size_milliseconds and inherited the millisecond bucket ladder. A quarter
of its samples landed above that ladder's top edge, and since Prometheus
returns the second-highest edge for a quantile in the `+Inf` bucket, its p95
panel showed a flat 5.00 kB rather than a measurement.
Adds beast::insight::Unit (Millis, Bytes) plus otelUnitCode(), carried on
EventImpl and selectable at makeEvent(). Naming the unit at creation is what
lets a backend pick the export unit and, through it, the bucket ladder.
- Collector gains a virtual makeEvent(name, Unit) whose default delegates to
the millisecond overload, so a collector that cannot act on a unit keeps
working unchanged. NullCollector and the Groups wrapper override it.
- The Groups override matters most: call sites reach a collector through a
Group, so forwarding only the prefixed name would silently drop the unit.
A test covers that hop specifically.
- Event gains notify(std::uint64_t) for non-duration samples, replacing
ServerHandler's `Event::value_type{response.size()}` -- wrapping a byte
count in a std::chrono::milliseconds compiles but reads as a duration to
everything downstream.
- EventImpl::value_type stays std::chrono::milliseconds. Widening it would
change the wire value of every existing StatsD timer, and metrics needing
finer resolution use the OTel-native microsecond instruments.
The StatsD collector deliberately keeps emitting `|ms`: that path is retired
here (its UDP port is commented out of the compose file and the integration
test fails if anything listens on 8125), so changing its wire format would
alter a legacy contract with no consumer and no way to verify it.
The exported name does not change yet -- OTelEventImpl still hardcodes its
unit. That follows with the unit-keyed histogram views.
clang-tidy runs misc-include-cleaner with WarningsAsErrors, so a symbol
reached only transitively fails CI. Add the direct includes for JLOG,
beast::Journal, StartUpType, TokenType, toBase58, std::exception and
std::size_t.
detachCallbacks() flips a flag that each observable callback checks on entry,
which leaves a callback already past that check running while the twelve
service stops below it tear down the state it reads. Stop the provider at the
same point instead: that joins the reader thread, so once it returns no
callback is running and none can start. Metrics recorded during the remaining
shutdown steps are no longer exported, which is the cost of the guarantee.
Build metricsRegistry_ in the member-init list rather than assigning it in
setup(). getMetricsRegistry() is read from the job queue and io threads, which
are running by then, so the later assignment was an unsynchronised write to the
handle those reads follow.
beast::insight instruments are created during ApplicationImp's member-init
list, and opentelemetry-cpp 1.28 never rebinds an already-vended Meter, so an
instrument created before the MeterProvider is published records nothing for
the rest of the process. Observable instruments carry the opposite constraint:
registering one arms the SDK reader thread, and its callbacks run hook handlers
that read services which do not exist that early.
Publish the provider in Telemetry's constructor, ahead of every producer, and
defer only the observables. Collector gains onCollectionReady() and
onCollectionStopping(); OTelCollector arms and disarms its gauges in response.
StatsDCollector starts its polling thread in its own constructor and had the
same hazard, so it uses the pair to gate that thread.
The metrics resource carries service.instance.id and is immutable once built,
so the node public key is resolved in Main.cpp, where a config error can still
be reported, and passed to makeApplication(). getNodeIdentity() remains
authoritative; both paths now share readNodeIdentity(), so telemetry cannot
report a key the node has abandoned.
An explicit ~ApplicationImp stops observing and stops telemetry, covering the
setup() failure paths that never reach run(). Telemetry::stop() is once-only
and no longer clears another instance's global pointer. The histogram view's
meter selector now matches the meter actually in use, so its bucket boundaries
apply for the first time.
makeTelemetrySetup() rejects a contradictory [telemetry] mutual-TLS
setup by throwing, but it is called from ApplicationImp's
member-initializer list. A try/catch in the constructor body cannot
reach a throw from there, and nothing further up the stack caught it
either, so a config mistake reached std::terminate: the default handler
printed a terminate dump and raised SIGABRT, leaving a core file
instead of a startup error.
Catch std::exception around makeApplication() in run(), report the
reason on stderr and return -1, so the failure is a clean non-zero exit
with a message an operator can act on. Only the construction is
wrapped. setup() starts subsystems whose shutdown order is delicate and
is left outside deliberately, because unwinding a half-started
Application would skip the normal stop sequence.
Gate both validation guards on enabled. A node with telemetry switched
off previously refused to start over certificate paths that nothing
would read.
Document both throws on makeTelemetrySetup(), state in
cfg/xrpld-example.cfg and the configuration reference that a partial
mutual-TLS setup is fatal and that the checks apply only when
enabled=1, and add a runbook troubleshooting entry keyed on the two
error messages.
Tests cover both guards with the message asserted so the two are told
apart, both enabled=0 paths, and the default plaintext configuration.
Node identity reached the OTel resource only as service.instance.id, which is
config-overridable and carries a deployment-chosen label rather than the node's
own identity. Add xrpl.node.id, set unconditionally from the node public key
(base58, TokenType::NodePublic), so traces and metrics share a stable per-node
key independent of [telemetry] service_instance_id.
Set on the tracer resource via Telemetry::setNodeId(), called from
ApplicationImp::setup() once nodeIdentity_ is known, and on the MetricsRegistry
resource via an added start() parameter. The beast::insight meter provider is
built in TelemetryImpl's constructor, before the wallet DB exists, so its
resource cannot carry the value; that path is left for later and the attribute
is omitted rather than stamped blank.
Also drops the transform/spanidentity collector processor added in
4a361a496d: per-node identity belongs on the resource, not copied onto every
span.
Nine conflicts, resolved as follows.
src/xrpld/app/ledger/detail/InboundLedger.cpp -- kept this branch's version.
phase10 sets the span's outcome/timeouts/peer_count attributes inline at each
exit; this branch replaced that with the idempotent finalizeAcquireSpan(), called
on all four exits (init, done, give-up, destructor). Taking phase10's blocks
would have set the outcome twice against a helper documented as not overwriting
what the real exit recorded. phase10's comment explains why peer_count must not
be read in a destructor; the helper solves that structurally by taking
std::optional<std::size_t> and being passed std::nullopt from there.
src/xrpld/telemetry/MetricsRegistry.cpp -- kept metric::ledgerEconomy over
phase10's "ledger_economy" literal. This branch added the naming check that
requires constants for converted families, so the literal would regress it. Took
phase10's comment cleanup.
src/xrpld/telemetry/MetricsRegistry.h -- kept registerRotationStateGauge(), which
only exists here, and took phase10's removal of the stale task-number comment.
validate_telemetry.py -- combined both. phase10 replaced serial metric polling
with a concurrent fan-out on one shared deadline, because 58 metrics x 45 s of
additive timeout overran the CI budget; that is kept. Its target list filters on
SKIPPED_METRIC_GROUPS rather than the two literals it hardcoded, so the
sync_diagnostics group stays owned by assert_sync_diagnostics_metrics() instead
of being polled and reported twice. Both SYNC_DIAGNOSTICS_GROUP and
METRIC_POLL_CONCURRENCY are needed and both are kept.
check_otel_naming.py -- both sides extend the rule docstring. Took phase10's
fuller Rule E text (doc discovery, allow-dotted markers) and re-appended rules
I/J/K/L, which exist only here.
expected_metrics.json -- the two sides add disjoint sibling groups, so both are
kept: sync_diagnostics alongside node_health_gauges, overlay_reduce_relay,
overlay_overflow, validation_lifetime_counters and not_asserted. Both dashboard
uids are kept, giving 16 asserted uids against 16 dashboards on disk.
expected_spans.json -- kept this branch's span set, a superset that adds the
acquire phase spans, ledger.serve, txset.acquire and peer.dial, and expands
ledger.acquire's required attributes. Took phase10's description, which documents
what the totals mean, and its note on how the RPC wildcard span is created.
total_span_types and total_unique_attributes are recomputed for the union: 48 and
74, since each side's figure counted only its own spans.
Docs: took phase10's more accurate wording on what the dashboard check actually
covers, and corrected the dashboard count from 15 to 16 where the merge made it
stale.
Verified: no conflict markers remain, both JSON contracts parse, both Python
files compile, asserted dashboard uids match the dashboards on disk exactly, and
the OTel naming check reports all layers consistent.
Four files conflicted.
xrpl.proto and TraceContextPropagator.h: phase-9's note is a semantic
superset of phase-8's and already carries no plan-folder pointer, so
phase-9's text is kept and nothing phase-8 said is lost.
node-health.json: the only change phase-8 makes to this file is renaming a
row that phase-9 had already deleted, so phase-9's file stands. Panel count
holds at 59 and both Validated Ledger Seq panels survive.
TESTING.md: resolved per hunk rather than by side. Phase-9's replacement of
the drifted span table and its new Grafana Cloud section are kept, and
phase-8's de-numbered "Test 3: Log-Trace Correlation" heading is taken --
keeping phase-9 wholesale would have reinstated the phase number that
phase-8 removed.
These comments were indexed by task, use-case and limitation numbers that
are defined only in planning documents outside the shipped tree. Nothing
in the repository defined them, so the cross-references resolved nowhere.
Each comment now states what the code does.
Two comment-only notes.
MetricsRegistry.cpp: record why the app and overlay includes exist and
what they cost. They are what makes loops.txt carry
xrpld.app <-> xrpld.telemetry and xrpld.overlay <-> xrpld.telemetry, where
ordering.txt previously had telemetry strictly below both. The observable
gauges are pull-model, so their callbacks need the concrete types to
sample live state. The cycle is confined to this translation unit: no
telemetry header includes app or overlay, and all of src/xrpld builds
into one target, so there is no header or link cycle. Inverting it needs
a metrics-source interface below overlay, which is left as follow-up.
Also note loops.txt is generated and must never be hand-edited.
docker-compose.yml: state the Compose >= 2.24.0 floor. The grafana
service uses the long-form env_file mapping, which older Compose cannot
parse, and it fails for the whole file rather than that one service. The
long form is required because .env.alerting is gitignored and absent in a
fresh clone, and the short form treats a missing env file as an error.
An InboundLedger destroyed while !isDone() recorded recordAbort() in the
metrics but left ledger.acquire carrying only the attributes set at
construction, so an abandoned acquisition was indistinguishable from one
still in flight.
Set outcome=aborted plus timeouts on that path, and add the val::aborted
constant. peer_count is deliberately omitted: reading it goes through
Overlay, and a destructor must not depend on Overlay still existing.
The status stays Unset for an abort, because InboundLedgers::stop()
clears every in-flight acquisition, so a clean shutdown would otherwise
report errors. done() now sets Error when failed_, which is an
unambiguous failure of the operation. Success is left Unset rather than
Ok, per the OpenTelemetry guidance that instrumentation should not
assert Ok.
rpcStart, jobQueue, jobStart and jobFinish each acquired a lock without
braces, so it stayed held to the end of the function and covered the OTel
recording calls this branch added. counters_.jobsMutex and
counters_.methodsMutex are process-wide, so every worker thread starting
or finishing a job serialised on the SDK's work.
That work is not a bare atomic add: each record builds a map-backed
attribute set and takes a spin lock inside the SDK, whose backoff reaches
a millisecond-scale sleep under contention.
Brace the lock plus the state it guards, then record after it releases.
None of the metric calls read lock-protected state, so this is
semantics-preserving, and rpcEnd() in the same file already had this
shape.
These comments pointed at a planning folder and at its rollout phase
numbering, neither of which is part of the shipped tree, so the
references would dangle for any reader of the repository. Each comment
now states the fact it was pointing at.
The reference docs had drifted from the code in ways that break the reader
rather than merely misinform: PromQL examples that return no data, a rollback
flag that is a no-op, a sampling knob that does not exist, and two span parents
that moved. Code is treated as the truth throughout; where the code is the
defective side, the doc now records it as a known issue instead of describing
the bug as intent.
Renames the docs missed: histogram names gain the exporter's unit suffix
(ios_latency_milliseconds_bucket and four siblings), ledger_history_mismatch
gains _total, the StatsD-era quantile label gives way to le buckets,
rpc.request becomes rpc.http_request, traces_spanmetrics_calls_total becomes
span_calls_total, and the nine dotted xrpl.* span attributes are recorded as
renamed rather than left as live keys.
Re-parenting: consensus.update_positions and consensus.check are children of
consensus.establish, not of consensus.round.
Units and labels: state_accounting_*_duration is microseconds, not seconds;
cache_metrics label values are case-sensitive; object_count carries demangled
C++ type names. Nodestore read and write latency stays microseconds -- the
nanosecond accumulator change did not move the exported unit.
Adds what shipped but was undocumented: the ledger.acquire span, seven
consensus.round events, twelve span attributes, node_writes_duration_us, the
7-day validation-agreement window, the TxQ admission and reduce-relay metric
families, metrics_endpoint, and the phase-10 validation workflow.
Corrects claims that never held: 10% head sampling (it is fixed at 100%),
configurable redaction (it is unconditional), -DXRPL_ENABLE_TELEMETRY=OFF
(the flag is -Dtelemetry=OFF, default ON), FindOpenTelemetry.cmake and the
xrpl_telemetry target (neither exists), Promtail and a StatsD exporter in the
pipeline (neither exists), and Loki stream selection on job= (only
service_name is a stream label).
Phase 9 is marked complete, its provisioned alerting is attributed to the
branch that shipped it, and Phase 11 stays at zero except the one prerequisite
its code closes. Counts are reconciled repo-wide: 41 emitted span families,
15 dashboards on disk with 14 asserted, 13 alert rules in 5 groups.
Hardens the gate that let this drift through: Rule E of the naming check now
covers the reference docs, its allow-dotted marker is key-scoped and warns on
stale or empty use, a missing checked file is reported instead of silently
skipped, the test suite runs in CI, and doc paths trigger the check.
C++ and CMake changes are comment-only: three MetricsRegistry instrument names,
eight OTelCollector claims of a metric-name prefix that formatName never adds,
and the telemetry option's inverted default.
