Files
rippled/.github/scripts/otel-naming/README.md
Pratik Mankawde 295ee1aa36 refactor(telemetry): name metrics with constants, and make CI require it (WP-A8)
Metric names and label keys were bare string literals, repeated across the
emit site, the gauge registration, the unit test, the workload manifest, the
dashboard queries and the reference table. A rename touched six places and a
typo in any one of them failed silently: a metric that never appears, or a
label that never joins.

The span side already had this right, with names and attribute keys declared
once in the *SpanNames.h headers and a CI rule rejecting literals at call
sites. That rule only ever covered spans, so the metric side had no
equivalent and no suffix convention was enforced by anything.

- Adds MetricNames.h declaring every instrument name, label key and bounded
  label value this story emits, grouped by subsystem, following the existing
  span-name header layout.
- Converts the call sites subsystem by subsystem. The emitted strings are
  unchanged: 75 names before, the same 75 after, verified by extracting the
  wire strings from both trees and diffing the sets.
- Extends the naming check with three rules: no literal instrument name or
  label key at an emit site, the duration and counter suffix conventions,
  and every name in the workload manifest resolving to a constant. The
  first rule is ratcheted per metric family so the pre-existing families
  warn rather than block, keeping the remaining work visible instead of
  forcing one unreviewable change.

Constants are character arrays rather than the span headers' StaticStr,
because the metrics API takes a string view that will not construct from it.

Two things the conversion exposed: a serve-refusal reason that the original
inventory missed because it is passed through a ternary, and a label whose
constant made it invisible to the checker's literal scan, which would have
failed a dashboard rule.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-27 12:20:01 +01:00

11 KiB

OTel naming-consistency check

check_otel_naming.py enforces the OpenTelemetry span-attribute naming convention documented in CONTRIBUTING.md across every layer of the telemetry pipeline. The *SpanNames.h constants are the single source of truth (L1); every other layer must agree with them.

Running locally

python .github/scripts/otel-naming/check_otel_naming.py

It takes no arguments, can be run from any directory inside the repo, and uses only the Python standard library (no pip install, matching the levelization check). A non-zero exit code means a violation was found; the output lists each violation as RULE | location | token | expected.

What it checks

The valid key set is derived dynamically from the OTel code — there is no hardcoded allowlist:

  • L1 keys come from the namespace attr { ... } blocks of every *SpanNames.h, resolving the makeStr("x") / join(seg::a, seg::b) DSL (cross-file, so join(seg::rpc, ...) resolves seg::rpc from the base SpanNames.h). Each constant is resolved against its own header, so two headers that define a same-named constant (e.g. a base attr::ledgerHash and a domain attr::ledgerHash) each contribute their real wire key — a later header cannot clobber an earlier one's value in a flat table.
  • Legitimate dotted keys = ONLY the keys the code actually sets as resource attributes, i.e. the entries inside Telemetry.cpp's Resource::Create({...}) call: the semconv::service::* keys (service.*) plus any attr::<name> constants passed there (xrpl.network.*). A dotted key that is declared in a header but never set as a resource attr is a span attribute in resource clothing — a Rule-A violation, even if it lives in the base SpanNames.h.
  • L1-metrics — instrument names, label keys and bounded label values come from the namespace metric / namespace label / namespace lval blocks of every *MetricNames.h, read as inline constexpr char NAME[] = "wire";. These headers deliberately do not use the makeStr/StaticStr DSL the span headers use: the OTel C++ API takes nostd::string_view, which constructs from char const* but has no constructor from std::string_view, so a StaticStr will not compile in an instrument-name or label-key position.

Rules (each fails the build, when its inputs are present)

Rule Check
A No stray dotted span-attribute key (only the derived resource keys may be dotted).
G Attribute keys are lower_snake_case (^[a-z][a-z0-9_]*$ per dot-segment) — no camelCase, UPPERCASE, or spaces.
F No string literals as attribute keys or span-name arguments in setAttribute/addEvent/span/rootSpan/childSpan (rootSpan shares span's (cat, prefix, name) signature). Attribute values are exempt (runtime data); *SpanNames.h definitions and test files are exempt.
B Every collector spanmetrics.dimensions name exists in the L1 key set.
C Every Tempo span-filter tag exists in the L1 key set.
D Every dashboard label resolves to an L1 span attribute, a native-metric label (L6, emitted by MetricsRegistry), or a Prometheus/Grafana builtin. TraceQL scope prefixes (span./resource./…) are stripped before the L1 lookup.
E No dotted xrpl.<domain>.<field> attribute key in the runbook (only the L1 resource attrs xrpl.network.* may be dotted). Span names, filenames, OTel-standard keys, and metric labels are not flagged.
I No string literals as metric instrument names or label keys — the mirror of Rule F. Applies to the name passed to an XRPL_METRIC_* macro or a meter->Create* factory and to the label keys in its label set. Label values, descriptions, *MetricNames.h, MetricMacros.h and test files are exempt. Scoped by metric family (first underscore segment): declaring a constant opts that family in, so the metric surface can be converted subsystem by subsystem. Unconverted families warn as Rule L.
J Metric instrument names follow the suffix conventions: lower_snake_case, no xrpld_/xrpl_ prefix (the exporter adds it), a counter ends _total, a histogram ends _us/_ms/_seconds, a gauge does not end _total. The instrument kind is read from the emit site, never guessed from words in the name — so a multi-series gauge carrying units in its label values (e.g. nodestore_latency observing write_mean_us) is not a violation.
K Every metric named in docker/telemetry/workload/expected_metrics.json resolves to a declared constant, so a rename in code cannot leave the workload validator asserting a name nothing emits. PromQL selectors (m{label="v"}) and exporter-appended histogram suffixes (_bucket/_count/_sum) are normalized away first; groups fed by another emit path (statsd_gauges, statsd_counters, spanmetrics) are out of scope by design.

Rule F runs unconditionally (it is a purely syntactic check on the call-sites and needs no *SpanNames.h), so a code path that calls SpanGuard::span/setAttribute directly without ever defining a header is still caught.

Warnings (printed, never fail the build)

Rule Check
H A namespace-qualified constant (e.g. foo::bar::myKey) used at a telemetry call-site is not defined in any *SpanNames.h. The constant should live in the proper header; defining it in-place bypasses rules A/G/F. Warns rather than fails — the argument may be a legitimately dynamic value, and the header may live on a later branch. Bare locals and std:: names are not warned.
L A literal metric name in a family that has no *MetricNames.h constants yet. Rule I's ratchet defers these instead of failing the build on the whole pre-existing metric surface at once; the warning keeps the outstanding conversion work visible rather than silently accepted.

Presence-gated

Every rule runs only when the source files it needs are present in the tree and is otherwise skipped (printed as SKIP: <rule> — <reason>), never failed. This keeps the check correct no matter how telemetry work is split across PRs — a stacked chain, one large PR, or independent per-stage PRs where (for example) the collector config lands before the dashboards. The collector/Tempo/dashboard/ runbook layers are introduced in later phases; on a branch without them, only the L1-intrinsic rules (A, G, F) run.