Skip to content

[Event Ledger] NVCA OTel Collector Extension for Event Ledger #809

Description

@shobham-nv

Description

FnDs (Event Ledger) needs per-instance lifecycle and failure detail for NVCF functions and tasks. Today the NVCA OTel sidecar already ships Pod events for container functions to FnDs. It does not ship NVCA ICMSRequest lifecycle/failure events, and it does not see helm pods in per-request sr-<uuid> namespaces.
This epic delivers a two-lane path into FnDs:

  1. Pod lane — kubelet/scheduler Events (container + helm + tasks), keeping the existing container Pod contract.
  2. ICMS lane — NVCA-emitted ICMSRequest Events with ledger annotations (icms_request_id, failure_category, etc.).
    Work spans NVCA (annotate + correlator), byoo-otel-collector (cluster-wide receiver + two lanes), and FnDs (kind-aware context). As an agreed SRE follow-up, add a read-only instance-summary API that joins Pod + ICMS on icms_request_id without changing ingest.

Definition of Done

  • NVCA stamps FnDs ledger annotations on ICMSRequest Events (functions + tasks), including failure_category on status updates
  • Multi-instance ICMS Event correlator preserves per-instance annotations under heartbeats
  • OTel collector watches cluster-wide and ships Pod and ICMSRequest lanes to POST /v3/ledger/k8s-events
  • Container Pod → FnDs path does not regress; helm pods and tasks are covered
  • FnDs uses kind-aware context (Pod unchanged; ICMSRequest includes icms_request_id)
  • Consumers can correlate lanes on icms_request_id (container and helm)
  • SRE instance-summary read API available (join on read; /stats and /events unchanged)
  • E2E scenarios pass for container, helm, and tasks (success + failure)

Resources (optional)

  • Bulleted list of links to related artifacts

By submitting this issue, you acknowledge that you are an assigned member of the NVCF development team and agree to follow our code of conduct and our contributing guidelines.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions