So let’s take a look at this graph, from the top down

  1. Queued at 2.97 sIt means that it takes 2.97s for the current request to be added to the request queue during the page loading process. Why does the call interface need to be queued according to the order of the queue? Because browsers limit the number of concurrent HTTP1.1 requests to the same Host at the same time, not all requests can be sent, so they need to be queued.

    The figure above shows the maximum number of concurrent HTTP requests each browser can make for the same domain name, so we Queued at 2.97sThe solution:

  • Increase the maximum number of concurrent requests and reduce the queuing time;

  • Different domain name calls, static resources and interface resources in different server domain name; Allocating resources to multiple hosts can increase the number of concurrent downloads, but increasing the number of parallel downloads will also increase the cost, which depends on bandwidth and CPU speed. Too many parallel downloads will reduce performance.

  • Reduce latency/blocking times, reduce proxy hops on the server, redirection, enable Keep Alive in requests, (temporarily) disable SSL, use HTTP instead of HTTPS;

  • Optimization process management, code execution level optimization, caching, DB, asynchronous, interface calls, etc.

  1. Started at 2.98s indicates the actual time to start processing after joining the queue. This and Queued at 2.97s are not included in the total time.

  2. Resource Scheduling    |    DURATION

    Queueing |  5.31 ms

    Resource scheduling, queue, represents the queue waiting time, is the time between adding to the queue to be processed and the actual start of processing.

  3. Connection Start | DURATION is the browser instructions to send the request, the request can wait time, generally is the proxy consultation, and wait for reusable TCP Connection release time, Do not include a DNS query, to establish a TCP connection wait time Stalled | 2.35 ms connection, stagnation of 2.35 ms solution portal _ reasons: _ Browsers have a limit on the number of concurrent connections to the same host domain, so if the current number of connections has been exceeded, other requests will be blocked, waiting for new connections, similar to the first Queued at; Scripts also block the download of other components; _ (1) The usual optimization includes: CSS, JS merge compression, image compression, Sprite diagram, static resources all on CDN; (2) Reasonable distribution of resources among multiple hosts can increase the number of concurrent downloads, but increasing the number of parallel downloads will also increase the cost, which depends on bandwidth and CPU speed, too much parallel downloads will degrade the performance; (3) The script is placed at the bottom of the page;

  4. Request/Response ** Request/Response ** Request/Response ** Request/Response ** Request/Response ** Request/Response ** Request/Response ** Request Sent 0.15ms Waiting(TTFB) 1.53s is about 1ms Waiting(TTFB) 1.53s is about 1ms Waiting(TTFB) 1.53s is about 1ms Waiting(TTFB) 1.53s is about 1ms Waiting(TTFB) 1.53s is about 1ms Waiting(TTFB) 1.53s Just like when you ask a question and your friend thinks for a while before giving you an answer, your friend’s Time To think is equivalent To TTFB (Time To First Byte). The less time your friend takes to think, the more intelligent or familiar he is with your problem. For the server, the shorter the TTFB time, the faster the server responds. It is usually less than 50ms. If the time is too long, it may be caused by the network, or the back-end query is slow, or the service performance is low, or the optimization is not done properly. Download time. This problem I looked up is the main solution to the manifold portal for you: **

  • 1. Front-end configuration solution

  • 2. The solution caused by the click event

  • 3, is the personal understanding of the returned JSON data is very large, this time may need to optimize the backend students

The principles and solutions to the problems have been posted, and the problems are so deep that it will take more time to fully understand them.

Recently encountered this problem, just want to record, may not have researched very thoroughly, great god many Pointers ~