Does any of APM customers in the Community have their entire APM infrastructure running on Windows 2008 Server platform? If so, please reach out to me.
We have performance issues in all of our environments (Prod, QA, & Dev) and running out of ideas.
Thanks in advance.
By entire infrastructure, do you mean including the agents? Or just the Enterprise Managers?
MoMs and Collectors (EMs) as far as infrastructure is concerned. Agent platform is out of scope for this.
In that case, I might be of use. We have our 18.104.22.168 EMs running on virtualized Windows Server 2008 R2. We have a variety of environments / setups:
1 MOM, 6 Collectors (5 Agent, 1 CEM)
1 MOM, 3 Collectors (2 Agent, 1 CEM)
1 MOM, 2 Collectors (Agent only)
3 Standalone EMs.
ok, cool, thanks jakbutler. So, have you come across performance issues on your APM infra? How big is your agent footprint per collector?
We've not had too many performance issues. We average about 1k metrics (a little over, depending on environment) per agent and have, in Production, around 5k agents at the moment; the number of metrics, and agents, will likely increase as we go through the various applications and enhance their instrumentation. Most of our issues have stemmed from the number of historical metrics reaching a clamp limit and similar. We recently had to add a collector to Production in order to ensure that we could handle a big spike in load.
What symptoms of performance problems are you having?