Skip to content

Conversation

@sc-zenokerr
Copy link
Contributor

@sc-zenokerr sc-zenokerr commented Dec 30, 2025

Fixes #3391

Question: I am not sure this is the right part of the document to add this. Where do you think is best?

image

@sc-zenokerr sc-zenokerr self-assigned this Dec 30, 2025
@sc-zenokerr sc-zenokerr requested a review from Frzk December 30, 2025 09:19
@Frzk
Copy link
Contributor

Frzk commented Jan 5, 2026

Question: I am not sure this is the right part of the document to add this. Where do you think is best?

To me, the whole "Scalingo for Caching" section should be reworked following the same structure as the one used for PG, MySQL or OpenSearch.
You'd then have a "Limitations" page where it would belong.

... But that would mean reworking the whole section, which is probably not a priority right now.

For now, I would add a h2 "Limitations" title at the bottom of the page and list it here.

@sc-zenokerr
Copy link
Contributor Author

Good call — I added a ## Limitations section at the bottom of the page and moved the HAProxy connection limit warning there to match the other database docs.

Copy link
Contributor

@Frzk Frzk left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A few nitpicks and it's OK for me :)
Thanks a lot <3

Comment on lines +442 to +450
{% warning %}
When using a Business plan, the HAProxy instances serving as entry points to your cluster have a default limit of **2,000 concurrent TCP connections**.

If this limit is reached, any new connection will be abruptly refused.

This can result in "Connection timed out" errors such as:

- _Redis::CannotConnectError : Connection timed out_
{% endwarning %}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd remove the warning block:

Suggested change
{% warning %}
When using a Business plan, the HAProxy instances serving as entry points to your cluster have a default limit of **2,000 concurrent TCP connections**.
If this limit is reached, any new connection will be abruptly refused.
This can result in "Connection timed out" errors such as:
- _Redis::CannotConnectError : Connection timed out_
{% endwarning %}
When using a Business plan, the HAProxy instances serving as entry points to your cluster have a default limit of **2,000 concurrent TCP connections**.
If this limit is reached, any new connection will be abruptly refused. This can result in errors such as:
```shell
Redis::CannotConnectError : Connection timed out

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Making that change results in the notice being mixed with the redis trademark notice:
image

I could add a trademark notice heading (which appears in the menu and search etc) like this:
image

Or something else. WDYT?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

BTW. I haven't pushed any changes... I'm looking for your opinion first!

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@benjaminach: I'm wondering if the notice is still mandatory... Any thoughts about it?

@sc-zenokerr sc-zenokerr requested a review from Frzk January 7, 2026 08:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Document our HAProxy connection limit in DB HA cluster

3 participants