TLS Advisory for AWS Users
DNS resolution & TLS certificate guidance for api.clickpost.in / www.clickpost.in
Who needs to act?
This advisory applies only if your integration with Clickpost (calls to api.clickpost.in, www.clickpost.in or any *.clickpost.in subdomain we've provided you) originates from inside an AWS VPC, and that VPC has an Interface VPC Endpoint for the execute-api service with Private DNS enabled — typically set up to support your own private API Gateway APIs.
If your integration calls Clickpost over the public internet (no VPC endpoint involved), no action is needed. If you're unsure whether this applies to you, check whether your AWS account has a VPC Endpoint for
com.amazonaws.
Symptom
Applications inside the affected VPC intermittently or consistently receive a TLS certificate mismatch when
calling our API domains.
The certificate presented is the default AWS *.execute-api
Root cause
Our custom domains sit on a regional AWS API Gateway, which is always backed by a target hostname of the form d-xxxxxxxxxx.execute-api.
When a VPC Endpoint for execute-api has Private DNS enabled, AWS's resolver logic in that VPC intercepts ALL DNS queries matching *.execute-api.
Recommended fix
The fix below was validated end-to-end with AWS Support on this exact issue and confirmed working. It
selectively bypasses the VPC Endpoint's private DNS override only for our specific API Gateway hostname, while leaving Private DNS intact for your own private APIs.
Step 1 — Use our API Gateway hostname
Use the following hostname in Steps 2 and 3 below. This is Clickpost's API Gateway wildcard domain and covers all of our subdomains (including api.clickpost.in and www.clickpost.in after the migration) — you do not need a separate hostname per domain or environment:
d-47wwl1sp68.execute-api.ap-south-1.amazonaws.com
Step 2 — Create a Route 53 Resolver Outbound Endpoint
- Place the endpoint in private subnets whose route table sends 0.0.0.0/0 to a NAT Gateway — not a public subnet routed directly to an Internet Gateway.
- Attach a security group that allows outbound UDP and TCP on port 53 to your chosen public DNS resolver (e.g. 1.1.1.1/32 or 8.8.8.8/32).
Step 3 — Create a Route 53 Resolver rule
- Rule type: Forward.
- Domain name: the exact hostname from Step 1 (avoid the wildcard *.execute-
api..amazonaws.com, so your own private APIs keep resolving via Private DNS as before). - Target IP: a public DNS resolver, e.g. 1.1.1.1 or 8.8.8.8.
- Associate the rule with the VPC(s) from which Clickpost APIs are called.
Step 4 — Validate
- From an instance inside the VPC, run dig or nslookup against the hostname from Step 1 and confirm it now returns a public IP, not a private (10.x/172.x) address.
- Re-run your integration call and confirm the TLS handshake now presents the *.clickpost.in certificate.
- Confirm your own private API Gateway APIs (if any) still resolve correctly via Private DNS — this rule is
scoped to a single hostname and should not affect them.
Common pitfall to avoid
Do not place the Resolver Outbound Endpoint in a public subnet routed straight to an Internet Gateway. Outbound Endpoints use private IP addresses, and an Internet Gateway cannot perform NAT for private. IPs — DNS queries to the forwarding target will simply time out. Always route the endpoint's subnet through a NAT Gateway.
Alternative (evaluate case-by-case)
If your VPC does not actually need Private DNS for its own private API Gateway APIs, disabling Private DNS on the execute-api VPC Endpoint entirely removes the conflict. This is more invasive since any internal service relying on that private DNS behaviour would need to switch to the VPC Endpoint's explicit DNS name instead — assess this against Step 2–4 based on your own environment before choosing it.
Need help?
Reach out to your Clickpost integration point of contact if you have any questions or need assistance validating the fix on your side.
Updated 3 months ago
