server/v0.1.9

beyond-pg server v0.1.9

fix(init): redial the substrate vsock instead of going silent forever.

An instd restart (every deploy:compute:local) left a postgres VM running but
permanently invisible to the host: the substrate thread exited on EOF and never
redialled, so instd saw no ready, no heartbeat replies, no PSI reports and no
workload logs, and marked the VM Degraded (guest disconnected). keep_alive now
distinguishes a disconnect (redial with capped backoff, re-sending
Ready{reconnect:true}) from an explicit Shutdown (stop for good). A first-connect
failure stays soft-fail, so Docker/no-AF_VSOCK environments are unchanged.

server/v0.1.8

perf(tuning): coalesce a scale-up into one retune, and stop sleeping …

server/v0.1.7

What's Changed

• fix(ci): sink release fires on server tags and breaks on slashed refs by @jaredLunde in #12

Full Changelog: server/v0.1.6...server/v0.1.7

server/v0.1.6

What's Changed

• feat(pgbouncer): peer so_reuseport workers to forward cross-worker query cancels by @jaredLunde in #10
• fix(tuning): retune shared_buffers behind a pgbouncer PAUSE by @jaredLunde in #11

Full Changelog: server/v0.1.5...server/v0.1.6

server/v0.1.5

beyond-pg v0.1.5: control-file extension filter + batched post_start

server/v0.1.4

beyond-pg v0.1.4: pre-baked PGDATA template (skip first-boot initdb)

server/v0.1.3

beyond-pg server/v0.1.3 — skip IPv6 DAD (nodad): fixes ~5s per-depend…

server/v0.1.2

beyond-pg v0.1.2: gate service.ready on pgbouncer :5432 accepting

server/v0.1.1

beyond-pg v0.1.1: emit ServiceKind::Ready so the postgres primitive c…

server/v0.1.0

beyond-pg + beyond-pg-init v0.1.0 (static musl)