ARCHIVE/rippled
1
0
Fork 0
mirror of https://github.com/XRPLF/rippled.git synced 2026-08-23 07:10:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(telemetry): address phase-9 metric gap-fill review comments

Browse Source 
- PerfLogImp::rpcEnd(): return after the requestId-not-found UNREACHABLE
  so a stale (now - epoch) duration is no longer recorded to the counter
  and histogram in release builds.
- MetricsRegistry peer-version gauge: compare versions numerically via
  BuildInfo::encodeSoftwareVersion() instead of a lexicographic string
  compare, stripping the non-digit prefix so peer 'rippled-X.Y.Z' lines
  up with our bare 'X.Y.Z'. Fixes every peer counting as higher-version.
- MetricsRegistry::stop(): call Shutdown() before ForceFlush() before
  reset() so the reader thread stops before teardown and no gauge
  callback fires during shutdown.
- MetricsRegistry::start(): extract initExporterAndProvider() and
  initSyncInstruments() helpers to keep each function under the line
  limit; no behavior change.
- time_in_current_state_seconds: read NetworkOPs::getServerStateDurationUs()
  (a lightweight accessor over StateAccounting) and convert microseconds
  to seconds, replacing the hardcoded 0.0.
This commit is contained in:
Pratik Mankawde
2026-07-07 20:47:05 +01:00
parent 6a2570e26e
commit 15f9c66dcf
5 changed files with 101 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
include/xrpl/server/NetworkOPs.h
View File

@@ -101,6 +101,13 @@ public:
[[nodiscard]] virtual OperatingMode
getOperatingMode() const = 0;
/** Time spent in the current operating mode so far, in microseconds.
Same value as `server_state_duration_us` in server_info; exposed as a
lightweight accessor so metrics can read it without building the full
server_info JSON on every collection tick.
*/
[[nodiscard]] virtual std::chrono::microseconds
getServerStateDurationUs() const = 0;
[[nodiscard]] virtual std::string
strOperatingMode(OperatingMode const mode, bool const admin = false) const = 0;
[[nodiscard]] virtual std::string

24
src/xrpld/app/misc/NetworkOPs.cpp
View File

@@ -276,6 +276,21 @@ class NetworkOPsImp final : public NetworkOPs
.start = start_,
.initialSyncUs = initialSyncUs_};
}
/**
* Time spent in the current operating mode so far. This is the same
* quantity reported as `server_state_duration_us` in json(): the
* elapsed time since the last state transition. Thread-safe.
*
* @return Duration since entering the current state, in microseconds.
*/
std::chrono::microseconds
currentStateDurationUs() const
{
std::scoped_lock const lock(mutex_);
return std::chrono::duration_cast<std::chrono::microseconds>(
std::chrono::steady_clock::now() - start_);
}
};
//! Server fees published on `server` subscription
@@ -358,6 +373,9 @@ public:
OperatingMode
getOperatingMode() const override;
std::chrono::microseconds
getServerStateDurationUs() const override;
std::string
strOperatingMode(OperatingMode const mode, bool const admin) const override;
@@ -981,6 +999,12 @@ NetworkOPsImp::getOperatingMode() const
return mode_;
}
std::chrono::microseconds
NetworkOPsImp::getServerStateDurationUs() const
{
return accounting_.currentStateDurationUs();
}
inline std::string
NetworkOPsImp::strOperatingMode(bool const admin /* = false */) const
{

4
src/xrpld/perflog/detail/PerfLogImp.cpp
View File

@@ -362,6 +362,10 @@ PerfLogImp::rpcEnd(std::string const& method, std::uint64_t const requestId, boo
{
// LCOV_EXCL_START
UNREACHABLE("xrpl::perf::PerfLogImp::rpcEnd : valid requestId input");
// Without an entry, startTime stays default-initialized; returning
// avoids recording a bogus (now - epoch) duration to the counters
// and OTel histogram below. UNREACHABLE does not halt release builds.
return;
// LCOV_EXCL_STOP
}
}

69
src/xrpld/telemetry/MetricsRegistry.cpp
View File

@@ -167,6 +167,22 @@ MetricsRegistry::start(std::string const& endpoint, std::string const& instanceI
JLOG(journal_.info()) << "MetricsRegistry: starting, endpoint=" << endpoint
<< ", instanceId=" << instanceId;
initExporterAndProvider(endpoint, instanceId);
initSyncInstruments();
registerAsyncGauges();
JLOG(journal_.info()) << "MetricsRegistry: started successfully";
#else
(void)endpoint;
(void)instanceId;
(void)enabled_;
#endif // XRPL_ENABLE_TELEMETRY
}
#ifdef XRPL_ENABLE_TELEMETRY
void
MetricsRegistry::initExporterAndProvider(std::string const& endpoint, std::string const& instanceId)
{
// Configure OTLP/HTTP metric exporter.
otlp_http::OtlpHttpMetricExporterOptions exporterOpts;
exporterOpts.url = endpoint;
@@ -205,9 +221,11 @@ MetricsRegistry::start(std::string const& endpoint, std::string const& instanceI
// Get a meter for all xrpld instruments.
meter_ = provider_->GetMeter("xrpld", "1.0.0");
}
// --- Create synchronous instruments ---
void
MetricsRegistry::initSyncInstruments()
{
// RPC per-method counters and histogram.
rpcStartedCounter_ = meter_->CreateUInt64Counter(
"xrpld_rpc_method_started_total", "Total RPC method calls started");
@@ -251,17 +269,8 @@ MetricsRegistry::start(std::string const& endpoint, std::string const& instanceI
// Note: xrpld_validation_agreements_total / xrpld_validation_missed_total
// are monotonic ObservableCounters created in registerValidationTotalsCounters()
// (below), observed from ValidationTracker's gross lifetime tallies.
// Register all observable (async) gauges.
registerAsyncGauges();
JLOG(journal_.info()) << "MetricsRegistry: started successfully";
#else
(void)endpoint;
(void)instanceId;
(void)enabled_;
#endif // XRPL_ENABLE_TELEMETRY
}
#endif // XRPL_ENABLE_TELEMETRY
void
MetricsRegistry::detachCallbacks() noexcept
@@ -288,9 +297,12 @@ MetricsRegistry::stop()
// to detach first.
callbacksDetached_.store(true, std::memory_order_release);
// Force-flush any pending metrics, then destroy the provider.
// This stops the PeriodicExportingMetricReader, which in turn
// stops invoking observable gauge callbacks.
// SDK teardown order: Shutdown() stops the PeriodicExportingMetricReader
// thread (so no further gauge callbacks fire) and performs the final
// collect-and-export drain itself. The trailing ForceFlush() is a
// redundant safety net (a no-op once the reader is shut down), then
// reset() destroys the provider.
provider_->Shutdown();
provider_->ForceFlush();
provider_.reset();
@@ -1018,7 +1030,20 @@ MetricsRegistry::registerPeerQualityGauge()
int higherVersionCount = 0;
int totalPeers = 0;
int divergedCount = 0;
auto const ownVersion = std::string(BuildInfo::getVersionString());
// Encode a version string into BuildInfo's comparable numeric
// form. Peers report the full "rippled-3.3.0-b0" string while
// our baseline is the bare "3.3.0-b0", and SemanticVersion
// requires a leading digit, so strip any non-digit prefix
// first. Numeric encoding avoids the lexicographic bug where
// "2.3.0" > "2.10.0" and "rippled-..." > "3...".
auto const encodeVersion = [](std::string_view v) -> std::uint64_t {
auto const firstDigit = v.find_first_of("0123456789");
if (firstDigit == std::string_view::npos)
return 0;
return BuildInfo::encodeSoftwareVersion(v.substr(firstDigit));
};
auto const ownEncoded = encodeVersion(BuildInfo::getVersionString());
app.getOverlay().foreach([&](std::shared_ptr<Peer> const& peer) {
++totalPeers;
@@ -1029,8 +1054,10 @@ MetricsRegistry::registerPeerQualityGauge()
}
if (pj.isMember(jss::version))
{
// Unparseable peer versions encode below ownEncoded, so
// the comparison correctly leaves them uncounted.
auto const pv = pj[jss::version].asString();
if (!pv.empty() && pv > ownVersion)
if (encodeVersion(pv) > ownEncoded)
++higherVersionCount;
}
// PeerImp::json() sets "track" to "diverged" when the peer's
@@ -1232,9 +1259,11 @@ MetricsRegistry::registerStateTrackingGauge()
}
observe("state_value", stateValue);
// TODO: Wire time_in_current_state_seconds to StateAccounting
// once a public accessor is available on NetworkOPs.
observe("time_in_current_state_seconds", 0.0);
// Time spent in the current operating mode, sourced from
// NetworkOPs' StateAccounting via a lightweight accessor.
auto const stateDurUs = app.getOPs().getServerStateDurationUs();
observe(
"time_in_current_state_seconds", static_cast<double>(stateDurUs.count()) / 1e6);
}
catch (...) // NOLINT(bugprone-empty-catch)
{

17
src/xrpld/telemetry/MetricsRegistry.h
View File

@@ -540,6 +540,23 @@ private:
opentelemetry::nostd::shared_ptr<opentelemetry::metrics::ObservableInstrument>
validationMissedObservable_;
/** Build the OTLP/HTTP exporter, periodic reader, resource attributes and
histogram views, then create the MeterProvider and meter. Extracted
from start() to keep each function under the 80-line limit.
@param endpoint OTLP/HTTP metrics endpoint URL.
@param instanceId service.instance.id resource attribute (may be empty).
*/
void
initExporterAndProvider(std::string const& endpoint, std::string const& instanceId);
/** Create the synchronous instruments (RPC and job-queue counters and
histograms, plus the external dashboard parity counters). Extracted
from start() to keep each function under the 80-line limit.
*/
void
initSyncInstruments();
/** Register all observable gauge callbacks with the OTel SDK.
Dispatches to one helper per metric domain so that each helper
stays well under the 80-line-per-function limit.