The workflow it schedules contains the action where we send the email via Sendgrid. We schedule the API workflow âscheduler_send_to_allâ from our admin panel.
It works fine in dev mode, but in run mode, nothing happens. No error appears in the server logs either. What am I missing?
Trying to run a workflow on a list of 50k Users is never going to work (at least reliably) - in your dev version itâs only 15 Users, so of course it will work there. (although I would still expect some of the Users to have received the email - so I donât know whatâs going on there)
Scheduling a workflow to run on a list works fine with a few things (usually only recommended for lists of maybe 10 things, 100 tops - it might even work with 1000 things - but 50k is way too much - at least on the current âcapacity basedâ pricing model, unless maybe youâre on a dedicated plan).
Besides, under the new pricing model it will be crazy expensive doing it this way, even if you could (paying for 100k outbound API calls, 100k Schedule actions, 100k Server actions)
You donât want to use Bubble to send individual emails for this - it will take forever and cost a fortune.
Use SendGridâs bulk email actions - you can send an email to 1000 email addresses with a single API call - so youâll just need to set up a recursive workflow to run 100 times (or until the full list has been sent) - so batch the contacts into groups of 1000.
I bulk email via SendGrid to 10s of thousands of users. Batching in groups of a thousand, in my experience, is even a bit much - not for SendGrid but for Bubbleâs backend. For example, ever try bulk deleting 1000 records within the data tab? Bubble times out unless youâre already paying for extra capacity. We batch in groups of 200 to 400 to have overhead available in the event someone wants to trigger additional (concurrent) mass email campaigns. Set this up in a recursive workflow, as @adamhholmes has suggested, and have logic in there to retry the current recursive loop (say up to three times) in the event it fails. This is what we do and it works well.