feat(semconv): add the ourios.miner.* metric registry + spec-first miner telemetry (§3.1.2/RFC0001.8/H2.2) - #160
Merged
Merged
Conversation
Author the §6.8 template-miner metric group in the weaver registry
alongside the compaction set, audited against the OpenTelemetry
semantic-conventions naming rules and the weaver registry policies:
- counters carry no `_total` suffix and singular `{annotation}` units
(`merges` `{merge}`, `parse_failures` `{failure}`,
`params.overflow` `{overflow}`, `template.version_changes`
`{change}`);
- fraction-of-total gauges use the conventional `.utilization` segment
with the dimensionless unit `1`
(`body_retention.utilization`, `params.overflow.utilization`),
not the non-conventional `.ratio`;
- elapsed-time histogram is `duration` (UCUM `s`), not `latency`;
- `confidence` is a histogram (unit `1`); `confidence.p50` / `.p01`
stay in-process named gauges per RFC0001.8 (the
histogram-vs-backend-quantile fork remains deferred);
- new attributes: `ourios.service` (the log's source service, distinct
from Ourios's own `service.name` resource attribute) and the
`ourios.miner.template_change` enum (widened / type_expanded, wire
values matching the audit `template_widened` / `template_type_expanded`).
`ourios.tenant` is `required`; `ourios.service` is `recommended`
(it may be absent on lines that carry no source service.name).
`weaver registry check -r semconv/registry --future` is clean; the
generated `ourios-semconv` is regenerated (generate + `cargo fmt` is a
no-diff). The exports test locks the new dotted names.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Owner
Author
|
@coderabbitai review |
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (5)
🚧 Files skipped from review as they are similar to previous changes (4)
📝 WalkthroughWalkthroughThis PR exposes the template-miner §6.8 OpenTelemetry metric set: it adds semconv metric/attribute definitions and constants, implements MinerMetrics with reservoir-backed p50/p01 gauges, wires metrics through the ingest/attach paths (including merge event_type), and replaces three stub tests with running telemetry validations. ChangesMiner OpenTelemetry Metrics
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
✅ Action performedReview finished.
|
Builds the §6.8 miner telemetry on the weaver-generated ourios_semconv constants (spec-first; the registry + generated module land in the prior commit). Retains observable-gauge handles, poison-safe collection-callback locks, and a pre-sized quantile reservoir. Flips §3.1.2 (mandatory metric set exposed), RFC0001.8 (confidence p50/p01 gauges), H2.2 (per-service params-overflow utilization alert). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
jensholdgaard
force-pushed
the
rfc0001-miner-semconv
branch
from
June 8, 2026 06:51
1dba5ff to
b76eca1
Compare
Contributor
There was a problem hiding this comment.
Pull request overview
Adds a spec-first OpenTelemetry semantic-conventions (weaver) registry for the RFC 0001 miner metrics and wires miner instrumentation/tests to consume the generated ourios-semconv constants (dotted ourios.miner.* names + ourios.* attribute keys).
Changes:
- Define
ourios.miner.*metrics and supporting attributes/enums insemconv/registry/*. - Regenerate and test
ourios-semconvexports for the new metric/attribute keys. - Implement miner telemetry (
crates/ourios-miner/src/metrics.rs), integrate it intoMinerCluster, and flip RFC red-gate stubs to in-memory OTel SDK tests.
Reviewed changes
Copilot reviewed 13 out of 14 changed files in this pull request and generated 9 comments.
Show a summary per file
| File | Description |
|---|---|
| semconv/registry/metrics.yaml | Adds the ourios.miner.* metric definitions (names, kinds, units, attributes). |
| semconv/registry/attributes.yaml | Adds ourios.service and ourios.miner.template_change attribute definitions. |
| docs/rfcs/0001-template-miner.md | Updates RFC text/table to reflect dotted-semconv naming (but still has some older name references elsewhere). |
| crates/ourios-semconv/tests/exports.rs | Extends exports test coverage to lock miner metric/attribute names to the registry. |
| crates/ourios-semconv/src/lib.rs | Adds generated constants for the miner metrics + new attribute keys. |
| crates/ourios-miner/tests/rfc_internal.rs | Implements RFC0001.8 test asserting p50/p01 gauges match in-process quantiles. |
| crates/ourios-miner/tests/invariants.rs | Implements §3.1.2 test asserting mandatory metric set appears at zero traffic. |
| crates/ourios-miner/tests/hazards.rs | Implements H2.2 per-service overflow-utilization alert-threshold test (minor message/name nits noted). |
| crates/ourios-miner/src/metrics.rs | New miner telemetry implementation (counters/histograms + observable gauges + init-seeding). |
| crates/ourios-miner/src/lib.rs | Exposes new internal metrics module. |
| crates/ourios-miner/src/cluster.rs | Integrates metrics into ingest/emit paths; records duration, confidence, overflow, parse failures, etc. |
| crates/ourios-miner/Cargo.toml | Adds opentelemetry + ourios-semconv; adds test-only ourios-telemetry, tokio, opentelemetry_sdk. |
| crates/ourios-core/src/audit.rs | Adds TemplateChange::event_type() for ourios.miner.template_change attribute emission. |
| Cargo.lock | Locks new dependency graph for added OTel + test crates. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/rfcs/0001-template-miner.md`:
- Around line 1850-1864: Rename the two metrics to follow the dotted-semconv:
change alias_assertions_total to ourios.miner.alias.assertions and
alias_retractions_total to ourios.miner.alias.retractions (remove the _total
suffix and add the ourios.miner alias namespace) in the table and update any
cross-references (e.g., H5, §6.7) or documentation mentions that refer to the
old names so they match the new metric identifiers.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: e7126aec-61e1-431e-b649-9d041f37a705
⛔ Files ignored due to path filters (1)
Cargo.lockis excluded by!**/*.lock
📒 Files selected for processing (13)
crates/ourios-core/src/audit.rscrates/ourios-miner/Cargo.tomlcrates/ourios-miner/src/cluster.rscrates/ourios-miner/src/lib.rscrates/ourios-miner/src/metrics.rscrates/ourios-miner/tests/hazards.rscrates/ourios-miner/tests/invariants.rscrates/ourios-miner/tests/rfc_internal.rscrates/ourios-semconv/src/lib.rscrates/ourios-semconv/tests/exports.rsdocs/rfcs/0001-template-miner.mdsemconv/registry/attributes.yamlsemconv/registry/metrics.yaml
…names Reconcile the pre-semconv flat metric/attribute names lingering in doc-comments, test panic/expect messages, and RFC 0001 §5 scenario text with the landed dotted-`ourios.miner.*` registry scheme: - metrics.rs: `parse_failures_total` / the `latency` histogram sentinel / `template_count` gauge prose → `ourios.miner.parse_failures` / `ourios.miner.duration` / `ourios.miner.template.count`. - tests/hazards.rs: H2.2 doc comment + the three gauge-read panic/expect messages → `ourios.miner.params.overflow.utilization` with `ourios.tenant` / `ourios.service` attributes. - RFC 0001 §5/§6.3/§6.4/§6.5/§6.7/§6.9: every remaining flat name (`template_count`, `merges_total`, `params_overflow_ratio`, `parse_failures_total`, `miner_latency_seconds`, the `confidence_p*` views, bare `tenant_id`/`service` attributes) reconciled to the dotted §6.8 table so the RFC is internally consistent with the generated `ourios-semconv` constants. Comment/doc/message only — no metric or attribute name, registry, or generated constant changed (weaver registry check stays clean). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
`service_of(&record.resource_attributes)` scans the resource attributes and allocates a fresh `String` every call. It was invoked from each hot-path helper (overflow retention, parse-failure, tokenizer-failure, emit), so a single ingested line re-scanned and re-allocated the service identity up to four times. Resolve it once at the top of `ingest` and thread it as `&str` through `ingest_string` / `ingest_structured` / `attach_and_maybe_widen` and the four helpers. Behaviour is identical — the same per-(tenant, service) instruments are driven with the same value — but the ingest hot path (pillar #2) now does one scan + one alloc per line. Also aligns the cluster.rs telemetry doc-comments touched by these helpers to the dotted `ourios.miner.*` names. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
`alias_assertions_total` / `alias_retractions_total` were the last two rows in the §6.8 table still carrying the Prometheus `_total` suffix and missing the `ourios.miner.` prefix, contradicting the amendment's own naming rule. Rename to `ourios.miner.alias.assertions` / `ourios.miner.alias.retractions` (and the §6.7 prose mention). These are future §6.7 metrics not yet in the registry, so this is a doc-only consistency fix. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add metric.ourios.miner.alias.assertions and metric.ourios.miner.alias.retractions to the weaver registry, both counters carrying the required ourios.tenant attribute, matching the RFC 0001 §6.8 dotted names. Regenerate ourios-semconv. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…* constants Replace the flat alias_assertions_total / alias_retractions_total metric names and the flat tenant_id attribute key with the generated ourios_semconv constants (OURIOS_MINER_ALIAS_ASSERTIONS / OURIOS_MINER_ALIAS_RETRACTIONS / OURIOS_TENANT), so the alias counters match the dotted RFC 0001 §6.8 names and are registry-backed. Add ourios-semconv as a dependency of ourios-core. Names only; the increments and tenant attribute value are unchanged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Review pass: the metrics-state write paths (record_line/overflow/ body_retention/set_template_count) now recover from a poisoned mutex like the collection callbacks (telemetry stays best-effort, never crashes the ingest hot path). The merges counter is init-seeded without the template_change attribute — that enum has no sentinel member, so a value appears only on a real merge (no out-of-contract time series). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
docs/rfcs/0001-template-miner.md (1)
1839-1842:⚠️ Potential issue | 🟠 Major | ⚡ Quick winMandatory zero-traffic guarantee conflicts with alias-counter behavior.
Line 1839 says every mandatory instrument is init-seeded for zero-traffic visibility, and Line 1851 includes alias counters in that mandatory set. But
crates/ourios-core/src/alias.rs(Line 205-209) explicitly documents those alias counters are not zero-seeded. Please align the contract (either seed alias counters or scope them out of the zero-traffic mandatory guarantee).Also applies to: 1851-1861
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/rfcs/0001-template-miner.md` around lines 1839 - 1842, Mismatch between the RFC's "init-seeded mandatory instruments" claim and alias.rs's doc that alias counters are not zero-seeded; reconcile by either (A) updating the RFC text around the "full mandatory set is exposed" / "init-seeded" language (and the table rows around the alias counters at the 1851–1861 section) to explicitly exclude alias counters from the zero-traffic mandatory guarantee, or (B) changing the alias implementation in alias.rs to perform initial seeding of alias counters at init (i.e., hook them into the same init-seed emission path used for other mandatory instruments) and update alias.rs documentation accordingly so both the RFC and alias.rs consistently state the same behavior. Ensure references to "alias counters" and "init-seeded/zero-seeded" are updated in both places.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@docs/rfcs/0001-template-miner.md`:
- Around line 1839-1842: Mismatch between the RFC's "init-seeded mandatory
instruments" claim and alias.rs's doc that alias counters are not zero-seeded;
reconcile by either (A) updating the RFC text around the "full mandatory set is
exposed" / "init-seeded" language (and the table rows around the alias counters
at the 1851–1861 section) to explicitly exclude alias counters from the
zero-traffic mandatory guarantee, or (B) changing the alias implementation in
alias.rs to perform initial seeding of alias counters at init (i.e., hook them
into the same init-seed emission path used for other mandatory instruments) and
update alias.rs documentation accordingly so both the RFC and alias.rs
consistently state the same behavior. Ensure references to "alias counters" and
"init-seeded/zero-seeded" are updated in both places.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: 0f3d558e-3151-41ed-80ca-b957b2129543
⛔ Files ignored due to path filters (1)
Cargo.lockis excluded by!**/*.lock
📒 Files selected for processing (13)
crates/ourios-core/Cargo.tomlcrates/ourios-core/src/alias.rscrates/ourios-core/src/audit.rscrates/ourios-miner/Cargo.tomlcrates/ourios-miner/src/cluster.rscrates/ourios-miner/src/lib.rscrates/ourios-miner/src/metrics.rscrates/ourios-miner/tests/hazards.rscrates/ourios-miner/tests/invariants.rscrates/ourios-miner/tests/rfc_internal.rscrates/ourios-semconv/src/lib.rsdocs/rfcs/0001-template-miner.mdsemconv/registry/metrics.yaml
✅ Files skipped from review due to trivial changes (2)
- crates/ourios-miner/src/lib.rs
- crates/ourios-semconv/src/lib.rs
🚧 Files skipped from review as they are similar to previous changes (8)
- crates/ourios-miner/Cargo.toml
- semconv/registry/metrics.yaml
- crates/ourios-miner/tests/invariants.rs
- crates/ourios-miner/src/metrics.rs
- crates/ourios-core/src/audit.rs
- crates/ourios-miner/tests/hazards.rs
- crates/ourios-miner/tests/rfc_internal.rs
- crates/ourios-miner/src/cluster.rs
Per review: replace the eight .unwrap_or_else(PoisonError::into_inner) call sites with a single lock_state() helper that matches on the LockResult (Ok guard / Err -> into_inner). Same poison-recovery behavior, expressed as an explicit match and DRYed to one place. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replace partial_cmp(...).expect("never NaN") with f64::total_cmp — a total
order with no panic on a telemetry hot path (no .expect() in non-test
code).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…nknown Drop the INIT_SENTINEL "__init__" attribute value: init-seeds are now attribute-less zero points, so the mandatory §6.8 set still surfaces at zero traffic (§3.1.2) without a sentinel that could collide with a real (unvalidated) tenant/service or violate the required template_change enum. Real series carry the registry-required attributes on real traffic. service_of now folds an empty service.name into SERVICE_UNKNOWN. Reservoir doc updated (dotted-semconv landed; the open fork is backend-derived quantiles, naming-independent). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…ion; cut per-line alloc service_of now returns Option<String> (None when service.name is absent or empty); ourios.service is recommended, so a service-less line is attributed to ourios.tenant alone rather than synthesized to "unknown". Drops the SERVICE_UNKNOWN sentinel. The per-tenant state splits named services (HashMap<String, _>, borrowed &str probe) from a dedicated no_service slot, so the common ingest path clones a key only on first sight of a (tenant, service) pair instead of every line. record_latency -> record_duration for grep-consistency with the histogram it records. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…nconditionally The module doc said "with no provider installed every record/add is a cheap no-op", implying the hot path is free when telemetry is disabled. The observable-gauge state updates (mutex + tally maps) run on every ingest regardless of the installed provider; document that this cost is intentional (in-process derived gauges) but bounded (clone-on-first-sight key, capped reservoir). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Contributor
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 15 out of 16 changed files in this pull request and generated 1 comment.
Comments suppressed due to low confidence (1)
docs/rfcs/0001-template-miner.md:1892
- The RFC text here says p50/p01 are recomputed on a background ticker and cached so export never blocks, and implies quantiles are evaluated over the histogram. The current implementation computes quantiles on collection from a bounded in-process reservoir (no ticker/caching). Please align the RFC with the implemented mechanism (or implement the ticker/caching described here).
named views derived from it in-process. The miner recomputes them
on a short ticker (default 10 s, configurable; the cost is one
quantile evaluation over the histogram per tenant per service per
tick — negligible relative to the hot path) and caches the value
between ticks so a metric export cycle never blocks on
The quantile gauge callback held the metrics Mutex while sorting each per-(tenant,service) reservoir (O(n log n) inside the tenant×service loop), stalling the ingest hot path during collection. Now it snapshots the sample windows under the lock (cheap O(n) copy) and sorts + computes the nearest-rank quantile via a free quantile_of() AFTER releasing the lock. Reservoir::quantile (now redundant) removed; tests use quantile_of(snapshot). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow-up to the off-lock quantile move: the test module references the free quantile_of via super:: (it was unqualified, breaking the lib-test build). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…ttribute Consistency with the dotted-semconv migration (the rest of the RFC + registry use ourios.tenant; wal_replay_progress is a future §6.9 WAL metric, documented not yet implemented). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The cluster "Not yet emitted" list was stale — three-zone confidence, type-expansion, reconstruction, overflow, and OTel exposition are all implemented; it now lists only the genuinely-pending items (Parquet records, WAL-backed sink, snapshot/recovery). The template.count registry brief now says tree leaves + structured-body templates, matching the exported count. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…gauges body_lines / body_retentions / template_counts updated via get_mut + insert-on-miss (bump_tenant / set_tenant helpers) instead of entry(tenant.clone()) / insert(tenant.clone()) — no per-line TenantId alloc on the ingest hot path for an already-seen tenant, matching the by_service path. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Contributor
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 15 out of 16 changed files in this pull request and generated 2 comments.
Comments suppressed due to low confidence (1)
docs/rfcs/0001-template-miner.md:1125
- The RFC pseudocode note says empty/whitespace-only input is "not a parse failure", but the current miner implementation treats
masked_strs.is_empty()as a parse failure (emits a lossy record, increments parse failures, and returnsNO_TEMPLATE). This should be reconciled so the spec matches the shipped behavior and the telemetry/tests (which use empty input to exerciseourios.miner.parse_failures).
# Note: an empty-after-whitespace string (the AnyValue
# carries `""` or only whitespace) is not a parse failure
# — it has zero tokens and the miner short-circuits with
# the cluster's `NO_TEMPLATE` sentinel rather than
# descending the tree. The pre-amendment cluster code
… split structure
Doc-only: the ServiceTally comment now names the dotted
ourios.miner.confidence.p50/.p01 gauges, and the by_service field doc
describes the ServiceTallies { by_name, no_service } split rather than a
stale inner Option<String> key.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This was referenced Jul 4, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Redo of #159 (closed for inverting the process) as a spec-first change: the weaver/semconv registry design is the main theme, authored and audited before any instrumentation. History reads spec-first — commit 1 is the registry + regenerated
ourios-semconv; commit 2 is the code built on the generated constants.1. OTel formulation rules applied (registry audit)
Authored against the OpenTelemetry semantic-conventions naming rules (
docs/general/naming.md,docs/general/metrics.md) and the weaver registry policies (weaver registry check --future, clean):_totalsuffix (Prometheus-ism; OTel reserves it for delta-backend monotonic sums). Names are bare nouns with singular UCUM{annotation}units —merges{merge},parse_failures{failure},params.overflow{overflow},template.version_changes{change}. The annotation matches grammatical number ({merge}, not{merges})..utilizationsegment (unit1), not a bespoke.ratio. OTel names a "fraction out of a total"utilizationwith the dimensionless unit1.duration(ourios.miner.duration, UCUMs) — the conventional segment for a discrete operation's elapsed time — notlatency.miner_latency_seconds→ourios.miner.duration+ units).stabilityandbrief.ourios.tenantisrequired(always present);ourios.serviceisrecommended(a line may carry no sourceservice.name) — an honest contract rather than over-promisingrequired.What changed vs the closed #159
#159 was treated as an un-audited starting reference. The audit changed three names:
ourios.miner.body_retention.ratioourios.miner.body_retention.utilization.utilizationis the OTel segment for a fraction-of-totalourios.miner.params.overflow.ratioourios.miner.params.overflow.utilizationourios.miner.latencyourios.miner.durationdurationis the OTel segment for operation elapsed timeThe
confidence.p50/confidence.p01gauges are kept as in-process named views of theconfidencehistogram (RFC0001.8); the histogram-vs-backend-quantile fork stays deferred to its own review, as the RFC requires.2. Registry entries (
semconv/registry/)Metrics group
ourios.miner.*(alongside the compaction set):merges(attrourios.miner.template_change),parse_failures,params.overflow,template.version_changesconfidence(unit1),duration(units)template.count,body_retention.utilization,params.overflow.utilization,confidence.p50,confidence.p01Attributes: reuse
ourios.tenant; addourios.service; add theourios.miner.template_changeenum (widened/type_expanded, wire valuestemplate_widened/template_type_expandedmatching the audit payloads).weaver registry check -r semconv/registry --futureis clean;weaver registry generate ... && cargo fmt -p ourios-semconvis a verified no-diff (the CI invariant).3. Generated constants (
ourios-semconv)Regenerated from the registry; consumed in code as
use ourios_semconv as semconv;— no flat/hand-written name strings anywhere. The exports test locks the new dotted names.4. Instrumentation built on the constants (
ourios-miner)crate::metricsinstruments the miner through the globalourios.minermeter (API-only dep per the §6.8 export-architecture split). Correctness baked in from the start (these were post-hoc review fixes on #159):ObservableGaugehandles are retained onMinerMetrics(a dropped handle deregisters its callback);lock().unwrap_or_else(PoisonError::into_inner)— never panic inside collect;VecDeque::with_capacity(RESERVOIR_CAP).Wired into
cluster.rs; the existing atomic counters feed the OTel instruments at the same sites. Three red-gate stubs flipped to AAA tests over anopentelemetry_sdkin-memory MeterProvider, asserting via the generated constants: §3.1.2 (mandatory set exposed at zero traffic), RFC0001.8 (p50/p01 gauges), H2.2 (per-(tenant,service) overflow utilization > 0.01 with sibling isolation).Invariants / hazards touched (CLAUDE.md §3 / §4)
template_changeenum keeps widening vs type-expansion distinct in the merges series; no silent merge introduced.params.overflowcounter +params.overflow.utilizationgauge are the per-service overflow signal; H2.2 asserts the >1% alert threshold per(tenant, service).template.version_changescounter tracks schema evolution.ourios.tenant.Verification (all green, reproduces CI)
weaver registry check -r semconv/registry --future— clean; generate + fmt — no-diffcargo test --all-features— 549 passed; 0 failed; 33 ignored (unrelated RFC red gates)cargo fmt --all --check— cleancargo clippy --all-targets --all-features -- -D warnings— cleanmdbook build— only the benign mdbook-mermaid version warning🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Tests
Documentation