Terraform "RequestError: send request failed" in CI

The provider tried to call a cloud API but the HTTP request never completed -- a dropped connection, DNS resolution failure, or TLS handshake error -- so no service-side response was received.

What this error means

apply/plan fails with "RequestError: send request failed" wrapping a network cause (connection reset, i/o timeout, no such host, EOF). It is typically transient and clears on retry from a healthy network.

terraform

Error: reading S3 Bucket (assets): operation error S3: HeadBucket,
exceeded maximum number of attempts, 3, RequestError: send request failed
caused by: Post "https://sts.amazonaws.com/": dial tcp: lookup sts.amazonaws.com:
no such host

Common causes

Transient network/DNS/TLS failure

A dropped connection, flaky DNS, or interrupted TLS handshake between the runner and the API endpoint aborts the request.

Egress restriction or proxy issue

A misconfigured proxy, firewall, or VPC endpoint can block or reset connections to the cloud API.

How to fix it

Retry, and confirm egress is healthy

Re-run the command; if it persists, verify the runner can resolve and reach the API endpoint.

Terminal

nslookup sts.amazonaws.com
curl -sS -o /dev/null -w '%{http_code}\n' https://sts.amazonaws.com/
terraform apply -input=false

Fix persistent connectivity

Confirm proxy/HTTPS_PROXY settings are correct for the runner.
Verify firewall/VPC-endpoint rules allow egress to the API hostnames.
Check DNS resolution from the runner if "no such host" recurs.

How to prevent it

Ensure runners have stable egress and DNS to cloud API endpoints.
Configure proxies/VPC endpoints correctly for provider traffic.
Run apply on a platform that auto-retries transient network failures.