lei.chen/myxce
1
0
Fork 1
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

docs(specs): record xce-node-manager interactions with service and provisioning specs

Browse source 
xce-node-manager: a "Cross-Spec Interactions" section (pending a
requirements revision; approved criteria unchanged) with six points: a
supervision/probe module shared with service-supervision, node-manager
components vs service jobs, a lease-based provisioner role, role
declaration and worker stop on decommission for provisioned hosts, a
host-loss grace period longer than worker restart plus rejoin, and a future
service proxy under the reverse-proxy role. design.md Open Questions points
to it.

The service-supervision, host-failure-job-recovery, elastic-hosts,
resource-provisioner and service-endpoint briefs cross-reference the
matching points; the roadmap lists the xce-node-manager revision.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Rongsong Shen 2026-10-10 07:46:24 +00:00
parent 922188b5d7
commit a68aac1a2b
Signed by untrusted user: rshen
GPG key ID: 80533536B0EB651C
8 changed files with 83 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

9
.kiro/specs/xce-node-manager/design.md
View file

@ -481,6 +481,15 @@ flowchart TB
- Validation checkpoint: after Phase 2, run the soak-harness scenario in Testing Strategy's E2E section before proceeding to Phase 3/4, since it is the closest thing to a direct regression test for bug-1760 itself.
## Open Questions / Risks
- **Cross-spec interactions (2026-10-10), pending a requirements revision:**
shared supervision/probe module with `service-supervision`; a lease-based
`provisioner` role for `resource-provisioner`; stopping the worker on
decommission and role declaration for provisioned hosts (`elastic-hosts`);
host-loss grace period longer than worker restart + rejoin
(`host-failure-job-recovery`); a future service proxy under the
`reverse-proxy` role (`service-endpoint`). See requirements.md,
"Cross-Spec Interactions".
- The exact Node Role Manifest file format (YAML to match `src/configs/*.yaml`'s existing convention, vs. a simpler key-value form) is left to implementation — an additive, appropriately implementation-level decision, not a design decision, since Data Models already fixes its logical shape.
- The precise heartbeat interval and "no progress" threshold (Requirements 4.2, 4.3) are implementation-time tuning decisions, following the same posture the superseded spec already took for its own liveness thresholds.
- `tools/appimage/entry.sh` bundles `haproxy` today with no dispatch case reaching it at all — this feature is what first wires it up (via `xce-node-manager` invoking it directly for the reverse-proxy role), which is a pre-existing gap this design closes as a side effect rather than a pre-existing behavior it must preserve.