What is High Availability?
High availability means that when a route, server, or node fails, the service can keep serving requests through automatic failover and redundancy, minimizing downtime.
Primary endpoint (recommended)
Suitable for most scenarios: more stable and better suited to production environments. Base URL:- High availability (built for production use)
- Behind Cloudflare CDN (more stable global access and protection)
- A 120-second forced timeout applies (imposed by Cloudflare CDN)
- If your request may exceed 120 seconds (for example, very long outputs, slow tool chains, or long-running streaming), we recommend switching to a secondary endpoint.
Secondary endpoints (no CDN, friendlier to long connections)
When you need longer request durations (for example, long-running streaming or complex tasks), you can use the secondary endpoints. Base URL (ready to copy):- No 120-second forced timeout (traffic does not go through a CDN)
- A theoretical request timeout limit of about 600 seconds (subject to your client and network environment)
- Not highly available: each is a single-point endpoint. If one is occasionally unreachable, you need to switch to the other secondary endpoint on the client side.
Recommendations
- Use the primary endpoint by default: more stable and better suited to production environments.
- If you clearly run into the 120-second timeout or need longer durations: switch to the secondary endpoints and implement failover on the client (switching between dc8 / dc15).
Notes
- You only need to configure the Base URL in the request address of the protocol you use. Different protocols only affect the authentication fields and request structure, not the domain choice described here.
- We recommend keeping the trailing
/consistent in your configuration (especially for the primary endpoint).