Maximizing TC agents performance

Hello. As sysadmin, I'm trying to deliver better performance of TC agents for my developers.

So what should I aim for?

  1. CPU generation
  2. CPU cores/threads 
  3. CPU frequency
  4. RAM capacity
  5. RAM frequency
  6. RAM bandwidth
  7. DISK bandwidth
  8. DISK latency

Currently I have 20 agents (10 win, 10 lin) deployed as VMs in ESXi on two servers (2xE5-2630 v4 + DDR4-2133 and 2xSilver 4114 + DDR4-2666 both with 10x240GB-SSD as JBOD). So each agent has 2c 8gb and its own ssd disk.

1 comment
Comment actions Permalink

In general, a better CPU, plentiful Memory for Java, and a fast Disk I/O will produce faster results, but not always. The agent hardware requirements are basically determined by the builds that are being run, so it depends on what the requirements of your builds will be and what would be optimal for each of those activities.

Running the TeamCity agent software does introduce a requirement for additional CPU time, but it can usually be neglected compared to the build process CPU requirements. Additionally, you can expect the build agent itself will consume about 500Mb of memory. The disk space required also corresponds to the disk usage by the builds running on the agent (sources checkouts, downloaded artifacts, the disk space consumed during the build; all that combined for the regularly occurring builds). It is also important to make sure the network bandwidth is not completely saturated during the checkout process and when downloading/uploading artifacts.

For additional information, please refer to our documentation here:


Please sign in to leave a comment.