Skip to content

chore(mq): nats-server logs about three Info lines per stream at boot, ~6,000 at 1,000 tenants #677

Description

@EricAndrechek

Area: mq / observability · gap · found via #612's "Follow-ups"

Expected: boot log volume does not scale with the tenant count.

Actual: the embedded broker gives each tenant its own JetStream streams (#612), and nats-server logs about three Info lines per stream at every boot — stream restore and consumer recovery — through the slogNATSLogger adapter (internal/mq/embedded.go:25-31, whose Noticef maps to slog). #612's own follow-up puts that at about 6,000 lines at 1,000 tenants, before WaveHouse logs anything of its own.

Impact: a boot at that scale buries WaveHouse's own startup lines and costs real money on a per-ingested-byte log backend. It also makes a genuine broker warning at boot hard to spot.

Note: the per-stream line count and the 1,000-tenant arithmetic are #612's measurement, quoted here; the adapter and the per-tenant stream layout are verified in the code.

Scope: the adapter is the single choke point — one option is to demote nats-server's Noticef to Debug, or to drop the known boot-time restore lines, and keep Warnf/Fatalf as they are.

Related: #612, #583


From the merged-PR follow-up sweep (#612 "Follow-ups"); validated by code-read against 4ff50745 on 2026-09-28.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    area/ingestIngest pipeline (Bento, batching, DLQ)area/observabilityMetrics, logs, traces, health, profilingchore

    Type

    No type

    Projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions