fix: surface OTel and sentinel visibility gaps

scripts/src/platform-infra-observability/actions.ts

@@ -120,6 +120,8 @@ export async function status(config: UniDeskConfig, options: CommonOptions): Pro
     return renderStatusTable({
       ok: result.exitCode === 0 && summary?.ready === true,
       target,
+      backendEndpoint: `${observability.traceBackend.serviceName}.${target.namespace}.svc.cluster.local:${observability.traceBackend.httpPort}`,
+      collectorEndpoint: `${observability.collector.serviceName}.${target.namespace}.svc.cluster.local:${observability.collector.otlp.grpcPort}`,
       summary,
       remote: parsed === null ? compactCapture(result, { full: false }) : null,
     });
@@ -143,6 +145,8 @@ export async function status(config: UniDeskConfig, options: CommonOptions): Pro
 function renderStatusTable(input: {
   ok: boolean;
   target: ObservabilityTarget;
+  backendEndpoint: string;
+  collectorEndpoint: string;
   summary: Record<string, unknown> | null;
   remote: Record<string, unknown> | null;
 }): RenderedCliResult {
@@ -167,6 +171,7 @@ function renderStatusTable(input: {
     `platform-infra observability status (${input.ok ? "ok" : "not-ok"})`,
     "",
     `target=${input.target.id} namespace=${input.target.namespace} ready=${textValue(input.summary?.ready)} route=${input.target.route}`,
+    `tempo=${input.backendEndpoint} collector=${input.collectorEndpoint} deployments=${deployments.length} pods=${pods.length} probes=${probes.length}`,
     "",
     "Deployments:",
     formatTable(["NAME", "READY", "AVAILABLE"], deployments.length > 0 ? deployments : [["-", "-", "-"]]),

scripts/src/platform-infra-observability/apply-status-scripts.ts

@@ -87,6 +87,7 @@ export function compactDiagnoseCodeAgentResult(value: unknown): Record<string, u
     spanCount: source.spanCount ?? null,
     services: source.services ?? null,
     servicePath: source.servicePath ?? null,
+    observabilityGap: source.observabilityGap ?? null,
     businessTraceIds: source.businessTraceIds ?? null,
     identity: compactDiagnoseIdentity(source.identity),
     agentrun: compactDiagnoseAgentRun(source.agentrun),

scripts/src/platform-infra-observability/diagnose-code-agent-script.ts

@@ -1986,8 +1986,26 @@ service_path = {
     service: ("reached" if service in services else "missing")
     for service in expected_services
 }
-service_path["complete"] = all(service in services for service in expected_services)
+missing_services = [service for service in expected_services if service not in services]
+service_path["complete"] = len(missing_services) == 0
+observability_gap = {
+    "status": "complete" if len(missing_services) == 0 else "missing-service-spans",
+    "expectedServices": expected_services,
+    "seenServices": sorted(services),
+    "missingServices": missing_services,
+    "complete": len(missing_services) == 0,
+}
+if missing_services and ("hwlab-cloud-api" in services or identity.get("runId") not in (None, "") or identity.get("commandId") not in (None, "")):
+    candidates.insert(0, {
+        "code": "observability_gap_missing_service_spans",
+        "label": "observability gap",
+        "confidence": 0.76,
+        "summary": "The business trace is correlated to Code Agent context but is missing expected service spans; do not interpret the missing manager/runner spans as proof that those services were not involved.",
+        "evidence": observability_gap,
+    })
 facts = []
+if missing_services:
+    facts.append("observability gap: missing service spans " + ",".join(missing_services))
 if http_summary.get("actorForbidden"):
     facts.append("actor forbidden")
 terminal_status = agentrun.get("terminalStatus")
@@ -2014,6 +2032,7 @@ summary = {
         "actorForbidden": http_summary.get("actorForbidden"),
         "terminalStatus": terminal_status,
         "failureKind": agentrun.get("failureKind"),
+        "observabilityGap": observability_gap.get("status"),
     },
 }
 evidence = {
@@ -2041,6 +2060,7 @@ payload = {
     "spanCount": len(spans),
     "services": services,
     "servicePath": service_path,
+    "observabilityGap": observability_gap,
     "businessTraceIds": business_trace_ids[:20],
     "identity": identity,
     "agentrun": {

scripts/src/platform-infra-observability/render.ts

@@ -613,6 +613,8 @@ export function renderDiagnoseCodeAgentTable(input: {
   const projectionLag = asPlainRecord(input.result.projectionLag);
   const summary = asPlainRecord(input.result.summary);
   const evidence = asPlainRecord(input.result.evidence);
+  const observabilityGap = asPlainRecord(input.result.observabilityGap);
+  const servicePath = asPlainRecord(input.result.servicePath);
   const rootCauses = asArray(input.result.rootCauseCandidates).map((item) => asPlainRecord(item) ?? {});
   const http = asPlainRecord(input.result.http);
   const services = joinValues(input.result.services, 54);
@@ -633,6 +635,14 @@ export function renderDiagnoseCodeAgentTable(input: {
     shortenEnd(textValue(candidate.summary ?? candidate.label), 80),
   ]);
   const httpRows = httpTableRows(http);
+  const expectedServices = asArray(observabilityGap?.expectedServices).map((item) => textValue(item)).filter((item) => item !== "-");
+  const seenServices = new Set(asArray(observabilityGap?.seenServices).map((item) => textValue(item)).filter((item) => item !== "-"));
+  const missingServices = new Set(asArray(observabilityGap?.missingServices).map((item) => textValue(item)).filter((item) => item !== "-"));
+  const serviceRows = expectedServices.map((service) => [
+    service,
+    textValue(servicePath?.[service] ?? (missingServices.has(service) ? "missing" : seenServices.has(service) ? "reached" : "-")),
+    missingServices.has(service) ? "missing" : seenServices.has(service) ? "seen" : "-",
+  ]);
   const queryClauses = asArray(input.query.queryClauses).map((item) => textValue(item)).filter((item) => item !== "-");
   const requestedRunId = textValue(input.query.runId);
   const requestedCommandId = textValue(input.query.commandId);
@@ -658,6 +668,10 @@ export function renderDiagnoseCodeAgentTable(input: {
     "Root causes:",
     formatTable(["CODE", "CONF", "SUMMARY"], rootRows.length > 0 ? rootRows : [["-", "-", "-"]]),
     "",
+    "Service trace coverage:",
+    formatTable(["SERVICE", "PATH", "SPAN"], serviceRows.length > 0 ? serviceRows : [["-", "-", "-"]]),
+    `  observabilityGap=${textValue(observabilityGap?.status)} missing=${joinValues(observabilityGap?.missingServices, 60)} seen=${joinValues(observabilityGap?.seenServices, 80)}`,
+    "",
     "HTTP:",
     formatTable(["METHOD", "ROUTE", "STATUS", "COUNT"], httpRows.length > 0 ? httpRows : [["-", "-", "-", "-"]]),
     "",