What are the current best practices? I you'd go with prometheus/grafana, how would you deploy it? Would you use a dedicated one for k8s?
Are you service-meshing? Would you use its own prometheus? You have a completely different setup, with other technologies?
This might be helpful to you -
<a href="https://signoz.io/blog/out-of-box-application-monitoring-prometheus/" rel="nofollow">https://signoz.io/blog/out-of-box-application-monitoring-pro...</a>
The following is working for me:<p>- dedicated prom instance for the cluster<p>- aggregated grafana with other prom instances<p>Thinking of investing in thanos in the near future.