As it turns out I have an application that has a wildcard subdomain. Our application is rather unique and we have a nearly infinite number of possible subdomains which could all produce unique content. Once the application has been running for a couple hours the metrics endpoint starts to time out for our Prometheus scraper.
I'm currently measuring on one of our nodes ~2900 unique domains known to PromEx which causes the metrics endpoint to take between 5-10 seconds to serve. The resulting output is 218873 lines long. That's a lot of metrics!
Would it be possible to disable the unique host tag being added to all the metrics? We don't need them per-host like this.
This bug is related to #183
As it turns out I have an application that has a wildcard subdomain. Our application is rather unique and we have a nearly infinite number of possible subdomains which could all produce unique content. Once the application has been running for a couple hours the metrics endpoint starts to time out for our Prometheus scraper.
I'm currently measuring on one of our nodes ~2900 unique domains known to PromEx which causes the metrics endpoint to take between 5-10 seconds to serve. The resulting output is 218873 lines long. That's a lot of metrics!
Would it be possible to disable the unique host tag being added to all the metrics? We don't need them per-host like this.