Chatwoot
Question about Performance #1358
Replies: 5 comments 14 replies
-
|
@anandsork, we handle a decent amount of traffic in our cloud version, which runs using the same code base. The performance aspect is mostly based on how you allocate the resources as your traffic starts growing. Initially, you could keep adding web/workers as the traffic increases. If you are facing a bottleneck at the database side, you could allocate more resources there. When you run into cases where you max out the resources you can allocate for a single server, you could look into implementing replicas to handle the reads and so ... So, Inherently there aren't any limits in the code that's stopping you from handling the scale. It's more about tweaking the architecture where chatwoot is running based on your growing requirements. If you face any specific issues, our team will be happy to help in resolving them. We also offer premium support solutions, where our engineers assist you with your self-hosted solutions, if that's something that interests you. |
Beta Was this translation helpful? Give feedback.
-
|
Yes. Horizontal scaling is possible. My question is more towards the limits it can reach to (or break) on a single server before we scale. Say 4GB memory machines (DB is on one instance and Other instance hold web+worker). |
Beta Was this translation helpful? Give feedback.
-
|
Makes sense. Unfortunately, we still need to do more performance testing in that aspect. So I don't have much metrics to share for the moment. |
Beta Was this translation helpful? Give feedback.
-
|
Hi @sojan-official related with this question, I've done performance test to send 1000 message from 10 diff numbers, and the chatwoot performance seems so slow while open the conversation list, cause chatwoot download all the conversation data which is thousand of chat data on that 10 conversations, why not load the chat when selecting the conversation ? I attached the snapshot about the requested timeframe btw
and this is the sample response data when loading the conversation list, and it's surprisingly big |
Beta Was this translation helpful? Give feedback.
-
|
Hi @pranavrajs i got this error after switch to develop branch and do Update dependenciesbundle Recompile the assetsrake assets:precompile RAILS_ENV=production Migrate the database schemaRAILS_ENV=production bundle exec rake db:migrate Restart the chatwoot serversystemctl restart chatwoot.target
I see on the log the issue related with web pack Is there any special config we need to compile but if I running it using |
Beta Was this translation helpful? Give feedback.
-
|
Please try this. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks @pranavrajs thanks it works, i spot the diff is on this line |
Beta Was this translation helpful? Give feedback.
-
|
@fiantinangon Precompile issues do not happen always. That's why we didn't add it to the upgrade steps. I will add it to the documentation. |
Beta Was this translation helpful? Give feedback.
-
|
Ok @pranavrajs i'll continue test the unread conversation performance and i'll get back asap about the result |
Beta Was this translation helpful? Give feedback.
-
|
We will be putting more efforts into load testing our critical endpoints using locust. |
Beta Was this translation helpful? Give feedback.
-
|
Hello! If I understand this right, we can safely increase the pod count of both web and worker servers, right? |
Beta Was this translation helpful? Give feedback.
-
|
That is correct. Just ensure that the database and Redis configs provided to these pods are the same. |
Beta Was this translation helpful? Give feedback.
-
|
Awesome, thanks! |
Beta Was this translation helpful? Give feedback.
-
This is not an issue, but a question.
How is the performance of the system for self-hosted solution? Can it handle 1000s of users at a time? The messages will reach to millions within 6 months of time. Any specific OS configuration needed for server & DB?
Beta Was this translation helpful? Give feedback.
All reactions