Self hosted : slow performances

Technical Help
,
1

Hello,

I use baserow for few week and I love this tool!

But I with my set up it’s a little bit slow. I mean when I want to fill a cell it’s fine, but for all others operations (create or delete a ligne, internal computation with formulas and look ups or the use of webhook and automation) the process often take more that 2 or 3s for each.

I hosted it on digital ocean and use 3 components:

*A database (1 vCPU / 2 GB RAM / 25 GB SSD - hosted in frankfurt)
*A docker to host the front (4 GB Memory / 80 GB Disk - hosted in frankfurt)
*An app to catch webhooks and make all automations (512 MB RAM | 1 vCPU - hosted in London, I’ll move it with the other soon).

I tried the cloud version and it’s much faster!!

Do you have any tips to improve the speed ?

Thanks,

2

I think that’s not enough if you want a fast bandwidth. I’ve been using contabo VPS S for self hosting and it’s super fast!!

image
image360×532 35.9 KB

PS: not doing a promotion

3

thanks for your reply @yashkumar

It seems more powerful and cheaper indeed!

PS: not doing a promotion

Let me know if you have a referal link, I’ll try it :wink:

Is 4bg of ram enough for the docker, from your point of view ?
Also, do you know if Contabo provide managed database ?

4

No more ideas ? Advices ?

5

Hey @Clapp sorry about the slow response, here are some questions/ideas:

  1. You mentioned you have 3 components. Which one of them is the baserow docker container as I wasn’t sure? Can you provide the exact anonymized command with env variables you are using to host Baserow itself and the server specs it is running on?

  2. Baserow is very database heavy, consider upgrading your database.

  3. Baserow itself needs at minimum 4GB of ram and also is quite demanding, consider upgrading the VM Baserow is running on.

  4. I can provide more specific steps if I know exactly how you are installing and running the Baserow server, but basically see the BASEROW_AMOUNT_OF_WORKERS and BASEROW_AMOUNT_OF_GUNICORN_WORKERS env variables from Configuring Baserow // Baserow

6

Also, be sure to double-check where this may come from (maybe just network latency or some other limitations rather than on Baserow’s side?).
If you have a powerful PC, you could try it on your own by accessing it locally if you don’t want to pay an extra fee for testing on DigitalOcean.

7

Thank you guys for your replies,

The problem definitively came ressources allowed to container and managed database.

I saw some bad comments about the managed database from Digital Ocean which can have poor performances:
https://laracasts.com/discuss/channels/servers/same-query-in-digital-ocean-managed-mysql-db-more-than-doubly-slow-than-on-droplet

So I’ve hosted the project elsewhere (OVH) with almost the double of ram on docker container (7gb) and the managed database (8gb) and it works super fast.

Thanks!

2 Likes