このページは、まだ日本語ではご利用いただけません。翻訳中です。
The following guide shows you how to set up the Post-function plugin to adjust request header names. In this example, we’ll edit two types of headers: headers set via a plugin (in this case, Rate Limiting), and latency headers from Kong Gateway).
- The Rate Limiting plugin returns headers such as
X-RateLimit-Remaining-<time>
andX-RateLimit-Limit-<time>
, where<time>
is the configured time span for the limit. - Kong Gateway adds latency headers to responses, such as
X-Kong-Upstream-Latency
andX-Kong-Proxy-Latency
. While you can turn these headers on or off inkong.conf
, they have fixed names that can’t be configured.
To change any header names, set up a Post-function plugin instance that runs in the header_filter
phase.
-
Create a Kong Gateway service:
curl -i -X POST http://localhost:8001/services/ \ --data "name=example-service" \ --data "url=https://httpbin.konghq.com/headers"
-
Add a route to the service:
curl -i -X POST http://localhost:8001/services/example-service/routes \ --data "name=test" \ --data "paths[]=/test"
-
Add a rate limiting plugin so that we can change its headers:
curl -X POST http://localhost:8001/services/example-service/plugins \ --data "name=rate-limiting" \ --data "config.second=5" \ --data "config.minute=30" \ --data "config.policy=local"
-
Create a Lua file for the plugin named
serverless.lua
(any name is fine). For example:return function() -- Rename rate-limit plugin headers -- X-RateLimit-Remaining-second: 1 -- X-RateLimit-Limit-second: 2 -- X-RateLimit-Limit-second changed to X-Rlls -- X-RateLimit-Remaining-second changed to X-Rlrs local kong_rl_headers = {} kong_rl_headers["x-ratelimit-limit-second"]="X-Rlls" kong_rl_headers["x-ratelimit-remaining-second"]="X-Rlrs" kong_rl_headers["x-ratelimit-limit-minute"]="X-Rllm" kong_rl_headers["x-ratelimit-remaining-minute"]="X-Rlrm" kong_rl_headers["x-ratelimit-limit-hour"]="X-Rllh" kong_rl_headers["x-ratelimit-remaining-hour"]="X-Rlrh" kong_rl_headers["x-ratelimit-limit-day"]="X-Rlld" kong_rl_headers["x-ratelimit-remaining-day"]="X-Rlrd" kong_rl_headers["x-ratelimit-limit-month"]="X-Rlln" kong_rl_headers["x-ratelimit-remaining-month"]="X-Rlrn" kong_rl_headers["x-ratelimit-limit-year"]="X-Rlly" kong_rl_headers["x-ratelimit-remaining-year"]="X-Rlry" local headers = kong.response.get_headers() for k, v in pairs(headers) do if kong_rl_headers[k] ~= nil then kong.response.set_header(kong_rl_headers[k], v) kong.response.clear_header(k) end end -- Add custom headers for latency kong.response.set_header("My-Custom-Proxy-Latency", ngx.ctx.KONG_PROXY_LATENCY) kong.response.set_header("My-Custom-Upstream-Latency", ngx.ctx.KONG_WAITING_TIME) end
-
Add the serverless function to the appropriate Kong entity by enabling the post-function plugin. For example, let’s apply it to the
test
route:curl -i -X POST https://localhost:8001/routes/test/plugins \ --form "name=post-function" \ --form "config.header_filter=@/tmp/serverless.lua"
-
Verify the changes. Make a call to the Admin API and check that the response header names have been changed:
curl -v http://localhost:8000/test
The response should show the new header names:
* Trying 127.0.0.1)... * TCP_NODELAY set * Connected to kong (127.0.0.1)) port 8000 (#0) > GET /test HTTP/1.1 > Host: localhost:8000 > User-Agent: curl/7.86.0 > Accept: */* > < HTTP/1.1 200 OK < Content-Type: application/json < Content-Length: 374 < Connection: keep-alive < Server: gunicorn/19.9.0 < Date: Mon, 27 Apr 2023 09:27:20 GMT < Access-Control-Allow-Origin: * < Access-Control-Allow-Credentials: true < X-Rlls: 2 < X-Rllm: 5 < X-Rlrm: 3 < X-Rlrs: 1 < My-Custom-Proxy-Latency: 1 < My-Custom-Upstream-Latency: 4
This is just a small demonstration of the power this plugin grants. You were able to dynamically inject Lua code into the header phase to dynamically transform the request without creating a custom plugin or reloading and redeploying Kong Gateway.