Going on 1.5 days no response on ticket for pro plan

Is this normal to be waiting 1.5 days with 0 reply on a support ticket regarding a domain on a pro plan?

The SLO for the Pro plan is 5 days - so yes, it could be normal for a ticket on a Pro plan zone to take more than 1.5 days.

https://support.cloudflare.com/hc/en-us/articles/200172476-Contacting-Cloudflare-Support#h_6SN2YW6wLVvk06Rbo1136r

2 Likes

Unfortunately, the time it takes for support to get back to you on the Pro plan is exceptionally long.
Usually, the community is the fastest way to solve problems. What error are you facing? What’s the ticket number?

2 Likes

I have had a load balancer setup for a few months now, has a pool of three nodes.

But for some reason it does not balance well. I have the % of traffic split across the three nodes but only two nodes are seeing traffic. At times all traffic gets sent to one node until it crashes.

It seems to be really unbalanced in directing the traffic and has been for months but worse as of late.

Ticket is: #2457784

Thank you.

1 Like

I believe we have seen similar reports in the past, I will escalate your ticket, and hopefully, you will get some replies quicker.

Try to give as much information as possible just so that you don’t waste support quotas (if they ask for more information, you will have to wait a few days for them to get back to you).

2 Likes

Thank you, I will see what else I can add to it.

2 Likes

Hi @wordpresseh

I would also recommend you check the content-type of your traffic split because not all requests are equal in load. For example if your running a WordPress site across your three nodes, then page requests would be processed by PHP and require a lot of CPU time in comparison to media requests which are read directly off the filesystem (little CPU overhead).

If you find that some nodes are sometimes getting overwhelmed with high CPU requests then it might be out of random chance that more dynamic page requests went to that node, draining the system resources, then subsequent requests are slower and the vicious cycle continues until that node just crashes.

If all your nodes have the same CPU and MEM specs and your load balancer split is equal then you can save yourself the cost of the load balancer service and just use DNS to split the traffic load randomly between the three nodes.

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.