docs: rewrote increasing request rate page

docs/increasing-request-rate.rst

@@ -4,25 +4,38 @@
 Increasing the request rate
 ===========================
 
-Increase the number of requests per second using a combination of the following steps:
+If you're not getting the desired/expected throughput there are a number of things you can do.
 
-#. Increase the number of users. To fully utilize your target system you may need a lot of simultaneous users, especially if each request takes a long time to complete.
+Concurrency
+-----------
 
-#. If response times are unexpectedly high and/or increasing as the number of users go up, then you have probably saturated the system you are testing and need to dig into why. This is not really a Locust problem, but here are some things you may want to check:
+Increase the number of Users. To fully utilize your target system you may need a lot of concurrent requests. Note that spawn rate/ramp up does not change peak load, it only changes how fast you get there. `Wait times <writing-a-locustfile.html#wait-time>`_ and sleeps *do* impact throughput. You can find a whole blog post on this topic `here <https://www.locust.cloud/blog/closed-vs-open-workload-models>`__.
 
-    -  resource utilization (e.g. CPU, memory & network. Check these metrics on the locust side as well)
-    -  configuration (e.g. max threads for your web server)
-    -  back end response times (e.g. DB)
-    -  client side DNS performance/flood protection (Locust will normally make at least one DNS Request per User)
+Load generation performance
+---------------------------
 
-#. If Locust prints a warning about high CPU usage (``WARNING/root: CPU usage above 90%! ...``) try the following:
+If Locust prints a warning about high CPU usage (``WARNING/root: CPU usage above 90%! ...``) try the following:
 
     -  Run Locust `distributed <https://docs.locust.io/en/stable/running-locust-distributed.html>`__ to utilize multiple cores & multiple machines
     -  Try switching to `FastHttpUser <https://docs.locust.io/en/stable/increase-performance.html#increase-performance>`__ to reduce CPU usage
     -  Check to see that there are no strange/infinite loops in your code
 
-#. If you are using a custom client (not HttpUser or FastHttpUser), make sure any client library you are using is gevent-friendly otherwise it will block the entire Python process (essentially limiting you to one user per worker)
+Also, if you are using a custom client (not HttpUser or FastHttpUser), make sure any client library you are using is `gevent-friendly <https://www.gevent.org/api/gevent.monkey.html>`__ otherwise it will block the entire Python process (essentially limiting you to one user per worker)
 
-.. note::
+If you're doing really high throughput or using a lot of bandwidth, you may also want to check out your network utilization and other OS level metrics.
 
-    Hatch rate/ramp up does not change peak load, it only changes how fast you get there.
+If you have issues with load generator performance and would rather pay to make it someone else's problem, you should check out  `Locust Cloud <https://locust.cloud/>`__.
+
+Actual issues with the system under test
+----------------------------------------
+
+If response times are high and/or increasing as the number of users go up, then you have probably saturated the system you are testing. This is not a Locust problem, but here are some things you may want to check:
+
+    -  resource utilization (e.g. CPU, memory & network)
+    -  configuration (e.g. max threads for your web server)
+    -  back end response times (e.g. DB)
+
+There are a few common pitfalls specific to load testing too:
+
+    -  load balancing (make sure locust isn't hitting only a few of your instances)
+    -  flood protection (sometimes a load test with the high amount of load from only a few machines will trigger this)