I’m currently running n8n on one single DigitalOcean droplet (4GB memory) with basically the default n8n setup. I have this first workflow (Workflow A) that sends out 200+ requests to Twilio. On the status callback of Workflow A, it calls another workflow (Workflow B) that simply listens to the request and does something else.
This is currently crashing my droplet whenever I trigger it, the execution logs show me that Workflow A ran up until the last Twilio node, as 90% of the end recipients received the Twilio messaging. Whereas Workflow B basically timed out on all incoming requests since the server crashed, and from my understanding of Twilio status callbacks, it consists of many callback requests for all status changes. I’m suspecting that the callbacks are overloading my local instance.
Without infrastructure changes, I’m thinking of adding a wait prior to hitting the Twilio endpoint. This is when I realized that everything in n8n gets wrapped up within a node before moving on to the next node, how can I throttle how I hit Twilio?