Fail fast when integration test environment is unavailable

[caydyan]

Summary

• add a preflight check before AVA integration tests that verifies every configured relay responds to /contacts
• add default HTTP timeouts for shared integration test request helpers
• document the readiness and per-request timeout environment variables

Fixes #415

Bounty payout BTC address: bc1qev5ant33v5y89qqjvcf4mh9hlax5svqf5xd7gc

Validation

• npx tsc --noEmit --pretty false
• git diff --check
• SPHINX_TEST_ENV_TIMEOUT_MS=500 npm test (expected fail-fast locally because relay ports 3001-3004 are not running; the command exits before AVA and reports ECONNREFUSED for the configured nodes)

[caydyan]

Current-head validation rerun on 92bae684eaf91569e50fea3729bd9cf8f628acd2:

• npx tsc --noEmit --pretty false passed.
• git diff --check origin/master...HEAD passed.
• SPHINX_TEST_ENV_TIMEOUT_MS=500 npm test exercised the new preflight and failed fast before AVA because the local integration relay nodes are unavailable (bob/alice/carol/dave/virtual nodes all returned ECONNREFUSED). That is the intended local smoke behavior for this branch: it reports the missing environment immediately instead of stalling in the integration suite.

Local worktree is clean after the rerun.

[caydyan]

Hi, just checking in on this after the validation rerun above.

Current status from my side:

• PR Fail fast when integration test environment is unavailable #815 implements fail-fast behavior for stalled integration-test relay startup.
• npx tsc --noEmit --pretty false passed on head 92bae684eaf91569e50fea3729bd9cf8f628acd2.
• git diff --check origin/master...HEAD passed.
• The local smoke path now exits with explicit unavailable relay-node diagnostics instead of hanging indefinitely.

If this satisfies issue #415's 300,000 sats bounty, BTC payout address is:

bc1qev5ant33v5y89qqjvcf4mh9hlax5svqf5xd7gc

Happy to rerun validation or adjust the behavior if you want a different timeout/error shape.