MetricsRegistry did two jobs. It owned the OTel metrics pipeline, and it
registered the observable gauges whose callbacks read live application
services. The second job is what made the whole class xrpld-tier, so the
pipeline's lifecycle -- the recording() gate and the stop() teardown that
closes a use-after-free window -- could not be unit-tested in xrpl_tests.
Split it in two:
- xrpl::telemetry::MetricsRegistry (libxrpl) owns the exporter, provider,
meter, the 16 synchronous instruments, recording(), stop(), and the
record*/increment* methods.
- xrpl::telemetry::AppMetricGauges (xrpld) owns the 19 observable gauges
and their callbacks, holding a reference to the core and to the
ServiceRegistry.
MetricMacros.h and ValidationTracker move with the core. The macros need
only recording() and meter(), both core members; the core holds a tracker
by value, and a libxrpl header cannot include one from src/.
ApplicationImp owns both objects and sequences them. The core is built in
the member-init list, so every synchronous instrument exists before any
subsystem can record one. The gauges are armed once overlay_ exists, the
last service their callbacks read. Shutdown detaches the gauge callbacks
before the core drops the provider, and each shutdown step is isolated so
a failure in one cannot skip the others.
That detach call is new. detachCallbacks() had no callers, and the flag it
sets is read by the gauge callbacks but can no longer be written by the
core, so the caller now has to make the ordering explicit.
The telemetry module links xrpl.libxrpl.core and xrpl.libxrpl.protocol
PUBLIC: ValidationTracker.h takes a LedgerIndex and MetricMacros.h takes a
ServiceRegistry, both in interfaces a consumer compiles against.
Adds a MetricsRegistry gtest that drives an enabled core with telemetry on
and pins the recording() gate, stop() leaving the registry inert, and
stop() being idempotent. The libxrpl test tree no longer depends on
xrpld.telemetry at all, and the two CMake workarounds that compiled xrpld
sources into xrpl_tests are gone.
Documentation and dashboard source links follow the code to their new
paths, split between the two classes by which one now defines each metric.
Three conflicts, all from both branches editing the same passage:
- Main.cpp: kept phase-9's wording. The metrics registry only exists here, so
"unwinding destroys little: the metrics registry, whose destructor joins its
export thread" is the true statement on this branch.
- TESTING.md: kept both paragraphs. They document different things (the
private [network_id], and the log path plus log_level).
- 05-configuration-reference.md: composed both. The identity is now resolved
before construction and never empty, so every producer stamps the node key
from the start; the only divergence left is a wallet that already holds a
different key, which corrects the tracer alone. Rewrote the earlier
"three producers" blockquote too: its "no fallback", "first boot ... left
off" and "Known issue" claims are what this change removes.
One silent break the merge could not flag: makeMetricsRegistryOptions() took
the std::optional<std::string> node key that used to be a constructor
parameter, and that parameter is now the resolved keypair. It takes the base58
string directly, and the constructor derives it from nodeIdentity_, which is
declared before both telemetry_ and metricsRegistry_.
resolveNodePublicKey() returned std::nullopt in three real cases: a first boot
with no wallet database, a standalone run (its wallet is a private temporary
database), and --newnodeid. Telemetry's resources are built during
ApplicationImp's member-init list and are immutable, so on those runs the node
reported an empty service.instance.id and no xrpl.node.id for the whole run,
while setup() minted a key moments later and patched only the tracer.
Replace it with resolveNodeIdentity(), which always returns a keypair: derived
from a configured seed, else read from an existing wallet database, else
minted. Main.cpp passes that pair to makeApplication(), ApplicationImp stores
it in nodeIdentity_ -- now declared before telemetry_ and no longer an optional,
because it is always set -- and builds the telemetry resource from it.
setup() calls getNodeIdentity(), which now persists rather than mints: it
stores the resolved pair when the wallet holds no identity, adopts the stored
one when it does, and clears first for --newnodeid. The write stays in setup()
because that is where the database exists; a standalone run has no persistent
wallet to write to, which is why the pair has to be decided before
construction rather than read back afterwards. Wallet gains storeNodeIdentity()
for that write, and getNodeIdentity(session) now uses it instead of repeating
the insert.
The three-argument makeApplication() mints a keypair, so jtx::Env and any other
test Application behave as a standalone run always did.
Also fold the three hand-rolled "meter from a NoopMeterProvider" copies into
telemetry::noopMeter(): the base-pointer call and the kMeterVersion argument are
both easy to get wrong alone, and the meter identity has to match the one the
histogram views select on.
The new gtest covers the wallet half: store-then-read, store not replacing an
existing identity, clear-then-store, and that the mint path persists. It adds
the tests.libxrpl > xrpl.rdb levelization edge, regenerated here.
One conflicted file, docs/telemetry-runbook.md, with three spots:
- Build section: both sides added different text at one point. Kept both,
incoming sentence first, then this branch's "Run against a live network".
- Disabling section, first spot: this branch's wording names the config
section and says no rebuild is needed, so it already covers the incoming
sentence.
- Disabling section, second spot: kept this branch's paragraph and folded in
the one point it lacked, that both flags have to be passed.
Three conflicts, all resolved by keeping this branch's rewrite and
re-applying the incoming change onto it:
- 09-data-collection-reference.md: phase-7 rewrote both attribute tables,
so the incoming table would have reverted them. Kept phase-7's and
re-applied the two "XRPL epoch" spellings.
- integration-test.sh: phase-7 moved these checks from StatsD to OTel and
no longer defines check_statsd_metric, so only this side compiles.
- TelemetryConfig.cpp: the incoming side carried networkTypeFromId(), which
this branch already has. Kept one definition and took the incoming
doc wording, which the auto-merged body below it already matches.
One conflict, in 06-implementation-phases.md: this branch had rewritten the
phase-4 task table with a Status column, a descoping note and a Spans Produced
section, while upstream corrected the class name in the old plain table. This
branch's section is kept and the name correction re-applied to its 4.1 row.
One conflict, in SpanGuard.h: this branch added struct TraceBytes and upstream
added enum SpanRole at the same position after TraceCategory. Unrelated
declarations, so both are kept.
Two conflicts, both resolved by composing the sides rather than taking one.
cfg/xrpld-example.cfg: this branch had moved the batch-processor keys under
their own heading while upstream edited them in place, so a merge-both would
have documented them twice. Upstream's range sentences are applied to the
relocated block and the head-sampling note keeps its position.
02-design-decisions.md: the summary table changed on both sides for different
reasons. Upstream renamed ledger_index to current_ledger_seq and ledger_seq;
this branch had corrected the PathFinding row to the keys it actually emits.
Both are kept.
Review feedback asked for a Histogram rather than a span attribute at these two
places. Both, not either: the attribute answers how big one sampled request was,
which an aggregate cannot, and the histogram answers the distribution across all
requests, which an unsampled trace never reveals. Both attributes stay.
The metrics land here rather than with the attributes because neither
HistogramBuckets.h nor the metric macro exists on the branch that added them.
Both use kObjectCountBuckets. The argument is the floor, not the ceiling: the SDK
default edges start 0,5,10,25, so an ordinary batch of one to five falls in a
single bucket and every quantile becomes an interpolation on one edge. The object
ladder puts five edges over the mass of both distributions. Path count is bounded
at 352 by kMaxPaths times kMaxAutoSrcCur and cannot saturate. Batch size can, at
roughly 333k, but no measured traffic goes near it, so the ladder is not widened
for a range nothing occupies; the runbook records the overflow query and a test
asserts it stays readable.
Review feedback on the RPC integration PR.
The childSpan examples could not work as written. childSpan() takes its parent
from the ambient context and uses impl_ only as a liveness gate, so an unscoped
SpanGuard parent produced two siblings rather than a parent and child. The parent
is now a ScopedSpanGuard, the child no longer reuses the parent's name, and the
examples pass a full dotted constant because childSpan() takes the name verbatim.
Five of the ten Rule D tests could not fail. Four passed an empty L1 key set,
which makes the rule skip validation altogether; the fifth asserted an empty
result against an escaped-quote selector that extracted no labels at all. Each
now passes a nonempty L1 set and carries a known-bad label in the same
expression, so it asserts both that the intended labels are accepted and that
Rule D ran. Verified by disabling the rule: the old tests stay green, the new
ones all fail.
Span kind is not fixed here. categoryToSpanKind and the span factories belong to
the telemetry library, so the role parameter is routed to that branch, and the
two call sites here follow once it exists.
These comments dated themselves against this branch or against an earlier
revision of the same change, neither of which survives a squash merge.
- 'as of this branch' in the pricing-case doc and the pass-through static_assert
becomes a statement about what the constants currently produce.
- 'Extracted from processGetObjectByHash()' and 'Split from start()' describe
edits internal to this change; both now say why the method stands alone.
- Recording.h: the mock-abstract mismatch is a standing consequence of a member
set that differs between builds, not something that 'has previously' happened.
- MetricsRegistry.cpp: describe the loops.txt entry as recording two cycles
rather than as what ordering.txt 'previously had'.
- InboundLedger.h: the acquire span is the only signal for back-fill cost; it
did not 'previously emit' nothing.
- check_bucket_parity.py: replace the eleven-phase drift story with the reason
the check exists, and point the failure message at HistogramBuckets.h and the
collector config instead of OpenTelemetryPlan/, which does not reach develop.
Comments and one error message only, no behaviour change.
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.
