实施新遗物会影响页面加载时间吗
does implementing new relic affect page load time
如果我们在 linux 服务器上安装新的 relic，是否会导致页面加载时间延迟？
我已经安装了新的 relic，它似乎导致页面加载时间有些延迟。
有人可以确认我吗？
提前致谢。
通过https://discuss.newrelic.com/t/will-newrelic-server-monitoring-affect-my-server-performance/4051:
We do extensive testing and estimate our overhead impact to be roughly
5% (+/- a few points) as it relates to CPU impact on the box and
application response time. This is based on a number of different
frameworks and load patterns we test with, as well as our default
instrumentation (byte code probes). Given that the Server Monitors do
much less instrumentation than our language agents, and mostly just
query data, we would expect this number to be even lower. Since every
application and server is different, we always encourage customers to
run their own tests in a pre-production environment to validate
overhead and let us know if they encounter anything different. Based
on our current traffic (datacenter bandwidth divided by number of
agents reporting), we calculate each agent consumes an average of
0.0008 Mb/s (megabytes/second) ~ 838 bytes/second, or roughly 50K bytes/minute.
所以是的，可能有可衡量的影响，但应该很小。
如果我们在 linux 服务器上安装新的 relic，是否会导致页面加载时间延迟？ 我已经安装了新的 relic，它似乎导致页面加载时间有些延迟。
有人可以确认我吗？
提前致谢。
通过https://discuss.newrelic.com/t/will-newrelic-server-monitoring-affect-my-server-performance/4051:
We do extensive testing and estimate our overhead impact to be roughly 5% (+/- a few points) as it relates to CPU impact on the box and application response time. This is based on a number of different frameworks and load patterns we test with, as well as our default instrumentation (byte code probes). Given that the Server Monitors do much less instrumentation than our language agents, and mostly just query data, we would expect this number to be even lower. Since every application and server is different, we always encourage customers to run their own tests in a pre-production environment to validate overhead and let us know if they encounter anything different. Based on our current traffic (datacenter bandwidth divided by number of agents reporting), we calculate each agent consumes an average of 0.0008 Mb/s (megabytes/second) ~ 838 bytes/second, or roughly 50K bytes/minute.
所以是的，可能有可衡量的影响，但应该很小。