Description of problem:
When exceeding haproxy_exporter_server_threshold (500° number of routes, all haproxy_server* metrics are not published anymore by haproxy


OpenShift release version:
* OCP 4.6
* OCP 4.7
* OCP 4.9.5

Cluster Platform:
* BM


How reproducible:
Always

Steps to Reproduce (in detail):
* deploy nginx sample application from catalog
* scrap metrics but using knowledge base article [^1] then grep haproxy_server metrics: `curl -kv -u  {USER}:{PASS} https://10.10.94.3:1936/metrics | grep haproxy_server`
* create 500+ routes : `for i in  $(seq 1 510) ; do echo route $i ; oc expose service nginx-sample --name nginx-sample${i} ; done`

[^1]:https://access.redhat.com/solutions/4998261

Actual results:
haproxy_server* metrics are not exposed

Expected results:
haproxy_server* routes are exposed

Impact of the problem:
impossible to monitor workload