Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- Route 53 checks reachable private endpoints from a health check source in a VPC. It normally needs subnets in three Availability Zones. Use two in US West (N. California).
- Verify the probe protocol, port, and path. FQDN probes resolve through the VPC's Route 53 Resolver.
- In staging, capture health state, the private DNS answer, and new client connection recovery separately. When both checked failover records are unhealthy, Route 53 still returns the primary.
Confirm that the source VPC can reach both endpoints
AWS announced Route 53 private endpoint health checks in all commercial Regions on November 19, 2026. Probes originate in a VPC you select and can check internal addresses or names that the VPC can reach. Choose the source VPC, Region, and subnets. Normally choose three subnets in distinct Availability Zones. AWS documents two subnets for US West (N. California).
Check the source subnet routes, security groups, and network ACLs for access to the endpoint port. Wait for the source to become Available before creating checks. Each source can support up to 200 private endpoint health checks. Keep the source, checks, and associated DNS records in the same AWS account.
Check the probe protocol and DNS resolution path
You can use a private IP address or fully qualified domain name with HTTP, HTTPS, or TCP. For HTTP or HTTPS, verify that a healthy endpoint returns an accepted 2xx or 3xx response within the probe's time limit. If you configure string matching, check the response content too. A TCP probe checks whether a connection can be established.
FQDN probes use the health check source VPC's Route 53 Resolver. AWS requires DNS resolution enabled and a private primary IPv4 CIDR in the VPC. For names only available through another private DNS setup, check the private hosted-zone association or Resolver forwarding rule, or probe by IP if appropriate. Use individual endpoint names for the two probes so neither targets the shared failover name.
Induce one primary failure and record three times
In staging, bind health checks to the primary and secondary records in a private hosted zone. Establish that both are healthy and a client in the intended VPC resolves the primary. Keep the secondary healthy. Make the controlled primary return HTTP 503 on its health path, or use an appropriate failure for the probe protocol.
Record separately when the primary check becomes unhealthy, when a representative client resolver begins to answer with the secondary, and when a new client connection completes a real request through it. Restore the primary and repeat the observations for health, DNS, and client behavior. You can also track existing connections separately from new ones.
Review the interval, caching, and failover rules separately
The standard check interval is 30 seconds, with a 10-second fast option. Route 53 marks the endpoint unhealthy after the configured failure threshold. DNS TTL, client caches, and reconnection behavior affect when an application uses the new address. Use the drill to measure recovery.
If both failover records have health checks and both are unhealthy, Route 53 returns the primary record. If the secondary lacks a check, Route 53 returns it when the primary is unhealthy even if the secondary service is unavailable. Record whether the secondary has a check when setting alert and drill expectations.
Sources
Frequently Asked Questions
Why does the client still use the old address after the primary is unhealthy?
Check the health state, current DNS answer, and client connection separately. TTL and connection reuse can keep a client on a prior address.
Will DNS return an empty answer if both endpoints are down?
For failover routing, Route 53 returns the primary when both records have health checks and both are unhealthy.
Can I replace subnets on an existing health check source?
The source VPC, subnets, and security groups cannot be edited after creation. Create a new source, reassociate the checks, and then remove the old source.