Welcome! Log In Create A New Profile

Advanced

[PATCH] Allow using nodelay=N semantics in limit_req configuration

Peter Shchuchkin
October 28, 2018 11:00AM
# HG changeset patch
# User Peter Shchuchkin <peters@yandex.ru>
# Date 1540737213 -10800
# Sun Oct 28 17:33:33 2018 +0300
# Node ID 70c0d476999d9b893c644606624134248ac7abad
# Parent 874d47ac871a4b62fbe0ff5d469a8ad7bc5a4160
Allow using nodelay=N semantics in limit_req configuration

This allows to use reasonably low limits while not forcing delay on normal
users.

In addition to standard "burst=A nodelay" form now the following form of
limit_req may be used:
burst=A nodelay=B, where B must be 0 <= B <= burst

burst=A nodelay=0 means the same as just "burst=A"
burst=A nodelay=A means the same as "burst=A nodelay"
burst=A nodelay=B means the first B requests matching limit_zone variable will
not be delayed and next requests will be delayed. The delay is calculated
against excess over B thus B+1 request will have effective excess=1.

When using limit_req with nodelay the responsibility of limiting requests speed
is on the client.
If client don't want or can't correctly limit its speed it will get 503 errors
and you will get numerous messages in error and access logs.
When using limit_req without nodelay, then every request that comes faster then
expected speed will be delayed. This is not always convenient. Sometimes you
want to allow normal client to make a bunch of requests as fast as possible
while still having configured limit on request speed.

Using this new semantics you can get the best from two worlds. Specifying
burst=A nodelay=B you allow clients to make B requests without any delay (and
without warnings in error log). If B requests are exceeded by client then
further requests are delayed, effectively limiting client rps to desired limit
without returning 503 errors. Thus one can ensure maximum speed for clients
with expected usage profile and limit all other clients to certain speed
without errors.

diff -r 874d47ac871a -r 70c0d476999d src/http/modules/ngx_http_limit_req_module.c
--- a/src/http/modules/ngx_http_limit_req_module.c Fri Oct 19 13:50:36 2018 +0800
+++ b/src/http/modules/ngx_http_limit_req_module.c Sun Oct 28 17:33:33 2018 +0300
@@ -499,12 +499,11 @@

excess = *ep;

- if (excess == 0 || (*limit)->nodelay) {
+ if (excess == 0 || ((ngx_uint_t) excess < (*limit)->nodelay)) {
max_delay = 0;
-
} else {
ctx = (*limit)->shm_zone->data;
- max_delay = excess * 1000 / ctx->rate;
+ max_delay = (excess - (*limit)->nodelay) * 1000 / ctx->rate;
}

while (n--) {
@@ -544,11 +543,16 @@

ctx->node = NULL;

- if (limits[n].nodelay) {
+ /*
+ * Delay when:
+ * excess > 0, nodelay = 0
+ * excess > 0, nodelay > 0, excess >= nodelay
+ */
+ if ((ngx_uint_t) excess < limits[n].nodelay) {
continue;
}

- delay = excess * 1000 / ctx->rate;
+ delay = (excess - limits[n].nodelay) * 1000 / ctx->rate;

if (delay > max_delay) {
max_delay = delay;
@@ -875,7 +879,7 @@
{
ngx_http_limit_req_conf_t *lrcf = conf;

- ngx_int_t burst;
+ ngx_int_t burst, nodelay_val;
ngx_str_t *value, s;
ngx_uint_t i, nodelay;
ngx_shm_zone_t *shm_zone;
@@ -885,6 +889,7 @@

shm_zone = NULL;
burst = 0;
+ nodelay_val = -1;
nodelay = 0;

for (i = 1; i < cf->args->nelts; i++) {
@@ -915,7 +920,20 @@
continue;
}

- if (ngx_strcmp(value[i].data, "nodelay") == 0) {
+ if (ngx_strncmp(value[i].data, "nodelay=", 8) == 0) {
+
+ nodelay_val = ngx_atoi(value[i].data + 8, value[i].len - 8);
+ nodelay = 1;
+ if (nodelay_val < 0) {
+ ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
+ "invalid nodelay \"%V\" - "
+ "must not be less then 0",
+ &value[i]);
+ return NGX_CONF_ERROR;
+ }
+
+ continue;
+ } else if (ngx_strcmp(value[i].data, "nodelay") == 0) {
nodelay = 1;
continue;
}
@@ -925,6 +943,23 @@
return NGX_CONF_ERROR;
}

+ if (nodelay) {
+ /* nodelay without explicit value */
+ if (nodelay_val < 0) {
+ nodelay_val = burst;
+ }
+
+ if (nodelay_val > burst) {
+ ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
+ "invalid nodelay value: %i - "
+ "must not be greater then burst: %i",
+ nodelay_val, burst);
+ return NGX_CONF_ERROR;
+ }
+ } else {
+ nodelay_val = 0;
+ }
+
if (shm_zone == NULL) {
ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
"\"%V\" must have \"zone\" parameter",
@@ -956,7 +991,7 @@

limit->shm_zone = shm_zone;
limit->burst = burst * 1000;
- limit->nodelay = nodelay;
+ limit->nodelay = nodelay_val * 1000;

return NGX_CONF_OK;
}
_______________________________________________
nginx-devel mailing list
nginx-devel@nginx.org
http://mailman.nginx.org/mailman/listinfo/nginx-devel
Subject Author Views Posted

[PATCH] Allow using nodelay=N semantics in limit_req configuration

Peter Shchuchkin 341 October 28, 2018 11:00AM

[PATCH] Tests: testing limit_req with "burst=A nodelay=B" configuration

Peter Shchuchkin 137 October 28, 2018 11:06AM

Re: [PATCH] Allow using nodelay=N semantics in limit_req configuration

Щучкин Петр 121 October 28, 2018 11:14AM

Re: [PATCH] Allow using nodelay=N semantics in limit_req configuration

Maxim Dounin 116 October 29, 2018 09:00AM

Re: [PATCH] Allow using nodelay=N semantics in limit_req configuration

Щучкин Петр 103 October 29, 2018 10:16AM

Re: [PATCH] Allow using nodelay=N semantics in limit_req configuration

Maxim Dounin 150 October 30, 2018 11:48AM

Re: [PATCH] Allow using nodelay=N semantics in limit_req configuration

Щучкин Петр 109 October 31, 2018 03:46AM

Re: [PATCH] Allow using nodelay=N semantics in limit_req configuration

Maxim Dounin 127 November 21, 2018 12:24PM



Sorry, you do not have permission to post/reply in this forum.

Online Users

Guests: 290
Record Number of Users: 8 on April 13, 2023
Record Number of Guests: 421 on December 02, 2018
Powered by nginx      Powered by FreeBSD      PHP Powered      Powered by MariaDB      ipv6 ready