The generated install command is host-generic, but nothing said so, and
cluster operators kept asking how to cover every member (#1816, #1618).
State it in the agent Quick Start next to the command generator pointer.
Refs #1816
Projection replay walked the entire alert event log on every boot, twice
(monitor start and the resource-store boundary), synchronously on the
serving path, and every replayed event queued a full incident-store JSON
rewrite. A 64k-event log made startup exceed the dev supervisor's
health-kill window, so the backend restarted forever and never served.
The event log now owns per-consumer projection watermarks in
alert_store_meta, WalkOldest takes an AfterID cursor, and the single
catch-up pass runs in the background at the canonical resource-store
boundary with periodic watermark checkpoints. The watermark only
advances when both the incident store and the canonical change recorder
are attached, so partial-surface passes repair without marking events
applied, and lowering the watermark forces a full repair replay.
Incident-store saves now coalesce: a burst of mutations queues one
whole-store serialization instead of one per event, live paths included.
Registers the alert-lifecycle-replay-startup-scalability coverage gap
and amends the alerts and monitoring subsystem contracts accordingly.
Restoring a configuration backup onto a Pulse instance with a new address
makes every agent vanish until each agent's configured --url (PULSE_URL for
containers) is updated, because agents dial the address they were installed
with. The post-migration checklist covered the token but not the address.
Refs #1813
Harden release workflow input transport and make native lifecycle proof honor each platform security boundary. Refresh stale telemetry and delivery qualifications.
Contract-Neutral: Formatting-only catalog updates and browser-verified alert copy; release compiler and native lifecycle changes update their owning contracts.
pulse-triage[bot]