One of the samples looks like this in the web console of OpenResty XRay. OpenResty XRay automatically identified those slow TCP connections and captured the TCP packets on them. In this case, we only want to see TCP connections with request latency longer than 100 milliseconds, for example. OpenResty XRay has the unique feature of capturing and analyzing network packets only on TCP connections that we are interested in. We can tell from the table above that all the slow requests are for the URI /view/shopcart/. We changed the Host column name so that we can share the sample here. OpenResty XRay also automatically collected the URI and other info about those slowest requests. We can notice that 5 requests with a long latency fall into the range of 131ms ~ 262ms. It then automatically creates the following logarithmic distribution chart for the response latencies. One of the great things about OpenResty XRay is that it can safely analyze online applications and servers without any changes in the running user processes. OpenResty XRay starts by sampling the requests that the web server processes (in this case, it is an OpenResty or Nginx web server) in the customer’s production environment. With so many possible suspects, dynamic tracing tools like OpenResty XRay can make it much easier to solve these tricky problems. The long latency issue can have many different causes, such as slow disk I/O, packet loss in the kernel network stack or other network devices, high CPU usage that slows down the applications or web servers, or just poor network links and connectivity. Some of the API requests take way too long, about 200 milliseconds, even though the application servers should handle them very fast. Try OpenResty XRay to analyze your running applications now
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |