Questions & Feedback regarding Percone Everest Scaling

Hey there. Amaaaazing project!!
I am wondering about scaling functionality. In the youtube video about scaling it is said that you can’t scale if you do not have the resources. I think this is really bad. All of our clusters use as few as possible resources and have autoscaler solutions in place to scale if more resources are being needed. It would be very important that this works. I think it is the biggest thing stopping us from further investigating on this.

A few more thoughts:

  • In an environment where one creates namepsaces on demand for e.x. tenants, I think it would be cruical that storage, pmm, etc. can be also not bound to a namespace, so that with dynamic namespaces this wouldn’t be an issue.

  • Backups: when we create a tenant, we use gitops to create namespace, s3 buckets, and other resources to setup everything. Is there anyway to have a process similar to creation of namespace, creation of everest resource database, creation of other gitops definitions?

  • Resource tagging would be amazing too. (Additional columns to search through, api endpoints to return tagged resources, etc.)

  • Resource logs that exatcly tell you when what happend because of what/who or finished happening.

  • Also, it would be amazing if you could prioritize namespaces or clusters to give their actions priority over e.x. development instances. (Simultaneous backups due to cron, order on when to go, as an example)

  • Can it be installed without cli, e.x. with gitops approach?

Thank you very much for all your effort. I am looking forward to this project. Much love!

Thank you so much for the detailed feedback!
We are very happy that you liked the Beta version and are very grateful for your suggestions for improvement.

I have passed everything on to the Everest team, we are very grateful.

I think someone from the team will comment and share ideas soon, probably something is already in the plans.

Hi @M11 ,

I think you may be misunderstanding this. This statement is true for anything. If I have 2 MySQL servers with 16GB RAM, 4vCPU, etc, I cannot scale without more resources. I can’t scale without A) more servers, or B) more resources within the existing servers.

Your autoscaler should add more resources automatically once you hit 70-80% utilization, so you have more for deploying. Alternatively, there should be an option to add more on the spot before deploying.

1 Like