Node.js HTTPS Agent keepAlive: Impact on Integration Test Performance Without Production Credentials

Node.js HTTPS Agent keepAlive in Tests

When running integration tests that hit an HTTPS endpoint without using production credentials, the default keepAlive setting on the https.Agent may influence both the number of TLS handshakes and the overall test duration. The Node.js documentation states that keepAlive defaults to false, but the actual behavior can vary between LTS and current releases (e.g., Node.js 12 through 18).

Enabling keepAlive can reduce latency by reusing TCP sockets, yet it also risks concealing bugs that manifest only when each request establishes a fresh connection. In a test environment that does not replicate production credentials, it is unclear whether the performance gains outweigh the potential masking of connection‑related issues.

Key constraints include:

• Testing must be performed against the target Node.js version in production.
• Connection teardown logic may differ when keepAlive is toggled.
• Test results should be reproducible across multiple runs.

Unresolved questions:

1. Does enabling keepAlive consistently reduce TLS handshakes across Node.js 12‑18 in integration tests?
2. Can keepAlive mask connection teardown bugs that only appear when each request uses a new socket?
3. Should integration tests mimic the production keepAlive setting, or should they run with keepAlive disabled to surface hidden issues?