After running some tests, I found myself compelled to use a Database Solution for creating a “Form Bot” on Telegram.
My main requirement is to have a bot with multiple types of forms, where users can fill out each form one by one. This cannot be achieved by adding just two triggers, as Telegram only allows one trigger. I also attempted to use the “Set” option, but it does not work when the SET data is cleared during the next execution. Therefore, the only solution now is to use a database to store the last message from the user and use it to switch to the other forms.
My question is: Is Airtable the fastest option available for me? (In terms of a No Code Database) I need to speed up my bot response. From what I can see, the Airtable process takes a bit longer, especially when there are multiple users chatting with the bot.
- Using N8N Self Hosted on Hetzner - Docker Through Coolify!
Thanks in advance.
Hi @anas, in my experience Airtable is relatively slow, both for search and write operations. I compared Baserow and Airtable performance recently by inserting 1,000 random records (with one text and one boolean field).
This operation took almost thrice as long on Airtable:
This was on a free plan, I don’t know if Airtable offers better performance on their paid plans (in which case the results could of course look different).
In addition, Baserow is powered by a regular PostgreSQL instance under the hood, so if you are self-hosting you’re not just limited to the n8n Baserow node (and the Baserow REST APIs), but could also connect to the PostgreSQL database directly for the best possible performance and flexibility through regular SQL queries.
For the best possible performance (while still having a UI letting you view and edit the database directly), Supabase would be king though. The above test (creating 1,000 records) takes less than a second using n8n:
So tl;dr Supabase is probably your best bet here, but Baserow is a little easier and will still be considerably better than Airtable.
Thanks for sharing your experience, @MutedJam . It’s really saving a lot of time. I just wanted to ask, did you test the self-hosted version of SupaBase or their free cloud plan?
Oh, good question! I was testing the free cloud plan from Supabase (but a self-hosted Baserow instance).
The comparison is also a bit unfair towards Baserow, as I could speed up the process big time by inserting stuff directly into the underlying Postgres instance on Baserow. So for the most accurate metrics you’d want to run your own benchmark emulating your actual use case.
The above is still a good first indicator I hope.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.