Running genesys cloud export flows to get JSON backups. The CLI hangs on the third flow and throws a 504 Gateway Timeout. The token has architect:flow:read scope. Is there a specific flag or pagination setting I’m missing? Here’s the output snippet:
> genesys cloud export flows --output ./flows
Exporting flows...
[ERROR] 504 Gateway Timeout on flow-id-abc123
The docs don’t mention rate limits for this command.
The 504 isn’t a rate limit issue. It’s the SDK trying to fetch the full flow definition in one go. Architect flows can get massive, and the default timeout for the underlying GET request gets hit before the JSON finishes parsing. The CLI export command doesn’t expose a timeout flag, which is annoying.
You’ll need to bypass the genesys cloud export command and hit the API directly with a Node script or curl. The trick is to request the flow definition with ?expand=flow but handle the response stream carefully, or just accept that you might need to increase the timeout in your HTTP client if you’re writing a custom script.
If you want a quick workaround without rewriting your backup strategy, try exporting individual flows by ID instead of the bulk command. It’s slower but more stable.
Here’s a quick Node.js snippet using the PureCloudPlatformClientV2 SDK to fetch a single flow without hitting the CLI’s hardcoded timeout. You can loop this over your flow IDs.
const platformClient = require('genesyscloud-purecloud-platform-client-v2');
async function exportSingleFlow(flowId, accessToken) {
const api = platformClient.FlowApi();
// Set a custom timeout if the SDK allows it,
// otherwise rely on the default which is usually 30s for SDK vs CLI's shorter limit
try {
// The 'expand' parameter is crucial here to get the full JSON structure
const response = await api.getFlow(flowId, { expand: 'flow' });
console.log(`Exported ${flowId}`);
return response.body;
} catch (error) {
if (error.code === 504) {
console.error(`Timeout on ${flowId}. Consider breaking down the flow or checking for circular references.`);
} else {
console.error(`Error exporting ${flowId}:`, error.message);
}
}
}
// Usage
// exportSingleFlow('your-flow-id-here', 'your-oauth-token');
If the flow is genuinely huge, check for any embedded complex conditions or large JSON data actions inside the flow. Those bloat the payload. Also, make sure your token has architect:flow:read and architect:flow:export if available, though read usually suffices for the GET endpoint.
The CLI export is convenient but brittle for large orgs. I’ve seen it fail on flows with over 500 steps consistently. Breaking it down manually is the only reliable way until they fix the stream handling in the CLI.