The question after every flag change is what happened to the errors, and normally that is two systems. Having both here means an agent can turn a flag on and read the traces that followed. That is also the reason to think about who holds this connection: 120 tools including flag creation is production control. Check the instance constraint early — federal and EU customers cannot use the hosted endpoint at all.
LaunchDarkly's hosted endpoint, spanning three product areas rather than one. There is an install wizard that configures your client for you.
- Feature management — creating and managing flags
- AgentControl — configuration management
- Observability — querying logs, traces, errors and dashboards
- Experiments and alerts
Browser OAuth; the client config needs no headers at all. Federal and EU instances are not supported by the hosted server — those use LaunchDarkly's local one.
Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen