chore: set Nomad server heartbeat_grace to 1m

[jakubno]

Increase the period for the client nodes before they're marked as dead

[cursor]

PR Summary

Low Risk
Low risk, but it changes Nomad server failure-detection timing, which could affect leader elections and cluster stability if the value is too strict for real-world network/CPU pauses.

Overview
Updates the Nomad server config generated by the AWS and GCP run-nomad.sh scripts to set server.heartbeat_grace = "1m", tightening the grace period used for server heartbeats.

^{Reviewed by Cursor Bugbot for commit 299ba78. Bugbot is set up for automated code reviews on this repo. Configure here.}

[claude]

LGTM — simple, consistent one-line config addition to both AWS and GCP Nomad server scripts.

Extended reasoning...

Overview

Adds heartbeat_grace = "1m" to the server {} block in both iac/provider-aws/nomad-cluster/scripts/run-nomad.sh and iac/provider-gcp/nomad-cluster/scripts/run-nomad.sh. The change is purely additive, symmetric across providers, and clearly scoped.

Security Risks

None. This is a Nomad cluster-internal timing parameter that controls how long the server waits before marking a node as down. No auth, crypto, or permission surfaces are touched.

Level of Scrutiny

Low. This is a single configuration knob with a well-defined semantics in Nomad. The value 1m is a relaxation of the default (10s), which makes the cluster more tolerant of transient network/CPU pauses before triggering leader re-election. The change is intentional and safe for operational stability.

Other Factors

No bugs were reported. The diff is minimal (one line per file), follows existing patterns exactly, and has no logical branching. No CODEOWNERS concerns for IaC config scripts.