http_server: Fix keep-alive for HTTP/1.1 requests #8761

iandrewt · 2024-04-25T11:12:55Z

This patch moves the initialisation of a HTTP/1.1 stream from the session initialisation to the session ingestion step.

Previously, using keep-alive on HTTP/1.1 would panic on the second request, as it would attempt to use components like the header map that are destroyed after each stream, but only initialised at the start of the session, when the HTTP protocol autodetection was run.

With this fix, HTTP/1.1 keep-alive is now enabled for the HTTP, OpenTelemetry, Elasticsearch and Splunk inputs

Fixes #8758

Enter [N/A] in the box, if an item is not applicable to your change.

Testing
Before we can approve your change; please submit the following in a comment:

[N/A] Example configuration file for the change
Debug log output from testing the change

Attached Valgrind output that shows no leaks or memory corruption was found

If this is a change to packaging of containers or native binaries then please confirm it works for all targets.

[N/A] Run local packaging test showing all targets (including any new ones) build.
[N/A] Set ok-package-test label to test for all targets (requires maintainer to do).

Documentation

[N/A] Documentation required for this feature

Backporting

Backport to latest stable release.

Fluent Bit is licensed under Apache 2.0, by submitting this pull request I understand that this code will be released under the terms of that license.

This patch moves the initialisation of a HTTP/1.1 stream from the session initialisation to the session ingestion step. Previously, using keep-alive on HTTP/1.1 would panic on the second request, as it would attempt to use components like the header map that are destroyed after each stream, but only initialised at the start of the session, when the HTTP protocol autodetection was run. Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_http plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_elasticsearch plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_opentelemetry plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_splunk plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

iandrewt · 2024-04-25T11:15:17Z

Valgrind output below

Ran the following curl command twice while fluent-bit was running

curl -v --http1.1 -d '{"hello":"world"}' -H 'Content-Type: application/json' 'http://localhost:9880/?[0-2]'

$ valgrind --leak-check=full ./bin/fluent-bit -i http -o stdout
==582166== Memcheck, a memory error detector
==582166== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
==582166== Using Valgrind-3.22.0 and LibVEX; rerun with -h for copyright info
==582166== Command: ./bin/fluent-bit -i http -o stdout
==582166== 
Fluent Bit v3.0.3
* Copyright (C) 2015-2024 The Fluent Bit Authors
* Fluent Bit is a CNCF sub-project under the umbrella of Fluentd
* https://fluentbit.io

___________.__                        __    __________.__  __          ________  
\_   _____/|  |  __ __   ____   _____/  |_  \______   \__|/  |_  ___  _\_____  \ 
 |    __)  |  | |  |  \_/ __ \ /    \   __\  |    |  _/  \   __\ \  \/ / _(__  < 
 |     \   |  |_|  |  /\  ___/|   |  \  |    |    |   \  ||  |    \   / /       \
 \___  /   |____/____/  \___  >___|  /__|    |______  /__||__|     \_/ /______  /
     \/                     \/     \/               \/                        \/ 

[2024/04/25 21:10:25] [ info] [fluent bit] version=3.0.3, commit=e47718ad23, pid=582166
[2024/04/25 21:10:25] [ info] [storage] ver=1.5.2, type=memory, sync=normal, checksum=off, max_chunks_up=128
[2024/04/25 21:10:25] [ info] [cmetrics] version=0.8.0
[2024/04/25 21:10:25] [ info] [ctraces ] version=0.5.0
[2024/04/25 21:10:25] [ info] [output:stdout:stdout.0] worker #0 started
[2024/04/25 21:10:25] [ info] [input:http:http.0] initializing
[2024/04/25 21:10:25] [ info] [input:http:http.0] storage_strategy='memory' (memory only)
[2024/04/25 21:10:25] [ info] [sp] stream processor started
[0] http.0: [[1714043453.692853667, {}], {"hello"=>"world"}]
[1] http.0: [[1714043453.723406781, {}], {"hello"=>"world"}]
[2] http.0: [[1714043453.724659850, {}], {"hello"=>"world"}]
[0] http.0: [[1714043461.957414623, {}], {"hello"=>"world"}]
[1] http.0: [[1714043461.958202834, {}], {"hello"=>"world"}]
[2] http.0: [[1714043461.958626946, {}], {"hello"=>"world"}]
[2024/04/25 21:11:03] [engine] caught signal (SIGINT)
[2024/04/25 21:11:03] [ warn] [engine] service will shutdown in max 5 seconds
[2024/04/25 21:11:04] [ info] [engine] service has stopped (0 pending tasks)
[2024/04/25 21:11:04] [ info] [output:stdout:stdout.0] thread worker #0 stopping...
[2024/04/25 21:11:04] [ info] [output:stdout:stdout.0] thread worker #0 stopped
==582166== 
==582166== HEAP SUMMARY:
==582166==     in use at exit: 0 bytes in 0 blocks
==582166==   total heap usage: 2,055 allocs, 2,055 frees, 2,205,819 bytes allocated
==582166== 
==582166== All heap blocks were freed -- no leaks are possible
==582166== 
==582166== For lists of detected and suppressed errors, rerun with: -s
==582166== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)

iandrewt · 2024-04-25T11:18:00Z

Debug output below

Ran the same curl request as the Valgrind check above

$ ./bin/fluent-bit -vv -i http -o stdout
Fluent Bit v3.0.3
* Copyright (C) 2015-2024 The Fluent Bit Authors
* Fluent Bit is a CNCF sub-project under the umbrella of Fluentd
* https://fluentbit.io

___________.__                        __    __________.__  __          ________  
\_   _____/|  |  __ __   ____   _____/  |_  \______   \__|/  |_  ___  _\_____  \ 
 |    __)  |  | |  |  \_/ __ \ /    \   __\  |    |  _/  \   __\ \  \/ / _(__  < 
 |     \   |  |_|  |  /\  ___/|   |  \  |    |    |   \  ||  |    \   / /       \
 \___  /   |____/____/  \___  >___|  /__|    |______  /__||__|     \_/ /______  /
     \/                     \/     \/               \/                        \/ 

[2024/04/25 21:15:48] [ info] Configuration:
[2024/04/25 21:15:48] [ info]  flush time     | 1.000000 seconds
[2024/04/25 21:15:48] [ info]  grace          | 5 seconds
[2024/04/25 21:15:48] [ info]  daemon         | 0
[2024/04/25 21:15:48] [ info] ___________
[2024/04/25 21:15:48] [ info]  inputs:
[2024/04/25 21:15:48] [ info]      http
[2024/04/25 21:15:48] [ info] ___________
[2024/04/25 21:15:48] [ info]  filters:
[2024/04/25 21:15:48] [ info] ___________
[2024/04/25 21:15:48] [ info]  outputs:
[2024/04/25 21:15:48] [ info]      stdout.0
[2024/04/25 21:15:48] [ info] ___________
[2024/04/25 21:15:48] [ info]  collectors:
[2024/04/25 21:15:48] [ info] [fluent bit] version=3.0.3, commit=e47718ad23, pid=583141
[2024/04/25 21:15:48] [debug] [engine] coroutine stack size: 24576 bytes (24.0K)
[2024/04/25 21:15:48] [ info] [storage] ver=1.5.2, type=memory, sync=normal, checksum=off, max_chunks_up=128
[2024/04/25 21:15:48] [ info] [cmetrics] version=0.8.0
[2024/04/25 21:15:48] [ info] [ctraces ] version=0.5.0
[2024/04/25 21:15:48] [ info] [input:http:http.0] initializing
[2024/04/25 21:15:48] [ info] [input:http:http.0] storage_strategy='memory' (memory only)
[2024/04/25 21:15:48] [debug] [http:http.0] created event channels: read=21 write=22
[2024/04/25 21:15:48] [debug] [downstream] listening on 0.0.0.0:9880
[2024/04/25 21:15:48] [debug] [stdout:stdout.0] created event channels: read=24 write=25
[2024/04/25 21:15:48] [ info] [sp] stream processor started
[2024/04/25 21:15:48] [ info] [output:stdout:stdout.0] worker #0 started
[2024/04/25 21:15:53] [trace] [io] connection OK
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] ret=149
[2024/04/25 21:15:53] [trace] [input chunk] update output instances with new chunk size diff=38, records=1, input=http.0
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_write] trying 43 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_write] ret=43 total=43/43
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] ret=149
[2024/04/25 21:15:53] [trace] [input chunk] update output instances with new chunk size diff=38, records=1, input=http.0
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_write] trying 43 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_write] ret=43 total=43/43
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] ret=149
[2024/04/25 21:15:53] [trace] [input chunk] update output instances with new chunk size diff=38, records=1, input=http.0
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_write] trying 43 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_write] ret=43 total=43/43
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:53] [trace] [io coro=(nil)] [net_read] ret=0
[2024/04/25 21:15:53] [trace] [downstream] destroy connection #42 to tcp://127.0.0.1:49556
[2024/04/25 21:15:54] [trace] [task 0x70ba88025220] created (id=0)
[2024/04/25 21:15:54] [debug] [task] created task=0x70ba88025220 id=0 OK
[2024/04/25 21:15:54] [debug] [output:stdout:stdout.0] task_id=0 assigned to thread #0
[0] http.0: [[1714043753.605323373, {}], {"hello"=>"world"}]
[1] http.0: [[1714043753.605581054, {}], {"hello"=>"world"}]
[2] http.0: [[1714043753.605736594, {}], {"hello"=>"world"}]
[2024/04/25 21:15:54] [debug] [out flush] cb_destroy coro_id=0
[2024/04/25 21:15:54] [trace] [coro] destroy coroutine=0x70ba80001050 data=0x70ba80001070
[2024/04/25 21:15:54] [trace] �[93m[engine] [task event]�[0m task_id=0 out_id=0 return=OK
[2024/04/25 21:15:54] [debug] [task] destroy task=0x70ba88025220 (task_id=0)
[2024/04/25 21:15:55] [trace] [io] connection OK
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] ret=149
[2024/04/25 21:15:55] [trace] [input chunk] update output instances with new chunk size diff=38, records=1, input=http.0
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_write] trying 43 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_write] ret=43 total=43/43
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] ret=149
[2024/04/25 21:15:55] [trace] [input chunk] update output instances with new chunk size diff=38, records=1, input=http.0
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_write] trying 43 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_write] ret=43 total=43/43
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] ret=149
[2024/04/25 21:15:55] [trace] [input chunk] update output instances with new chunk size diff=38, records=1, input=http.0
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_write] trying 43 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_write] ret=43 total=43/43
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] try up to 1024 bytes
[2024/04/25 21:15:55] [trace] [io coro=(nil)] [net_read] ret=0
[2024/04/25 21:15:55] [trace] [downstream] destroy connection #42 to tcp://127.0.0.1:49570
[2024/04/25 21:15:56] [trace] [task 0x70ba88025ec0] created (id=0)
[2024/04/25 21:15:56] [debug] [task] created task=0x70ba88025ec0 id=0 OK
[0] http.0: [[1714043755.870506601, {}], {"hello"=>"world"}]
[1] http.0: [[1714043755.871465391, {}], {"hello"=>"world"}]
[2] http.0: [[1714043755.871616262, {}], {"hello"=>"world"}]
[2024/04/25 21:15:56] [debug] [output:stdout:stdout.0] task_id=0 assigned to thread #0
[2024/04/25 21:15:56] [debug] [out flush] cb_destroy coro_id=1
[2024/04/25 21:15:56] [trace] [coro] destroy coroutine=0x70ba80008370 data=0x70ba80008390
[2024/04/25 21:15:56] [trace] �[93m[engine] [task event]�[0m task_id=0 out_id=0 return=OK
[2024/04/25 21:15:56] [debug] [task] destroy task=0x70ba88025ec0 (task_id=0)
[2024/04/25 21:15:59] [engine] caught signal (SIGINT)
[2024/04/25 21:15:59] [trace] [engine] flush enqueued data
[2024/04/25 21:15:59] [ warn] [engine] service will shutdown in max 5 seconds
[2024/04/25 21:16:00] [ info] [engine] service has stopped (0 pending tasks)
[2024/04/25 21:16:00] [ info] [output:stdout:stdout.0] thread worker #0 stopping...
[2024/04/25 21:16:00] [ info] [output:stdout:stdout.0] thread worker #0 stopped

iandrewt · 2024-04-25T11:19:22Z

And finally for reference, here's the curl output for one execution of above

$ curl -v --http1.1 -d '{"hello":"world"}' -H 'Content-Type: application/json' 'http://localhost:9880/?[0-2]'
* Host localhost:9880 was resolved.
* IPv6: ::1
* IPv4: 127.0.0.1
*   Trying [::1]:9880...
* connect to ::1 port 9880 from ::1 port 40718 failed: Connection refused
*   Trying 127.0.0.1:9880...
* Connected to localhost (127.0.0.1) port 9880
> POST /?0 HTTP/1.1
> Host: localhost:9880
> User-Agent: curl/8.7.1
> Accept: */*
> Content-Type: application/json
> Content-Length: 17
> 
* upload completely sent off: 17 bytes
< HTTP/1.1 201 Created
< content-length: 0
< 
* Connection #0 to host localhost left intact
* Found bundle for host: 0x643f13fd0220 [serially]
* Re-using existing connection with host localhost
> POST /?1 HTTP/1.1
> Host: localhost:9880
> User-Agent: curl/8.7.1
> Accept: */*
> Content-Type: application/json
> Content-Length: 17
> 
* upload completely sent off: 17 bytes
< HTTP/1.1 201 Created
< content-length: 0
< 
* Connection #0 to host localhost left intact
* Found bundle for host: 0x643f13fd0220 [serially]
* Re-using existing connection with host localhost
> POST /?2 HTTP/1.1
> Host: localhost:9880
> User-Agent: curl/8.7.1
> Accept: */*
> Content-Type: application/json
> Content-Length: 17
> 
* upload completely sent off: 17 bytes
< HTTP/1.1 201 Created
< content-length: 0
< 
* Connection #0 to host localhost left intact

iandrewt · 2024-04-25T23:30:07Z

Not sure why that macOS test failed - concerning since the specific failure relates to in_http, but I just tested it on my M1 Pro (I don't have any x86_64 Macs I can easily set up a test on) with the CI repo locally running FLB_OPT="-DFLB_JEMALLOC=Off" ./ci/scripts/run-unit-tests.sh and the tests passed

iandrewt · 2024-04-26T00:41:17Z

As for the performance difference with this change, I'm seeing about a 25% improvement with this setup:

/etc/nginx/conf.d/default.conf

upstream fluent {
  server 127.0.0.1:9880;
  keepalive 16;
}
server {
  listen 80;
  location / {
    proxy_pass http://fluent;
    proxy_http_version 1.1;
    proxy_set_header "Connection" "";
  }
}

script.js

import http from 'k6/http';
export const options = {
  vus: 100,
  duration: '30s'
};
export default function() {
  const url = 'http://localhost';
  const payload = '{"hello":"world"}';
  const params = {
    headers: {
      'Content-Type': 'application/json'
    }
  };
  http.post(url, payload, params);
}

fluent-bit -i http -o null

Before:

$ k6 run script.js

          /\      |‾‾| /‾‾/   /‾‾/   
     /\  /  \     |  |/  /   /  /    
    /  \/    \    |     (   /   ‾‾\  
   /          \   |  |\  \ |  (‾)  | 
  / __________ \  |__| \__\ \_____/ .io

     execution: local
        script: script.js
        output: -

     scenarios: (100.00%) 1 scenario, 100 max VUs, 1m0s max duration (incl. graceful stop):
              * default: 100 looping VUs for 30s (gracefulStop: 30s)


     data_received..................: 163 MB  5.4 MB/s
     data_sent......................: 165 MB  5.5 MB/s
     http_req_blocked...............: avg=2.45µs  min=250ns   med=801ns  max=37.58ms p(90)=1.21µs  p(95)=1.38µs
     http_req_connecting............: avg=488ns   min=0s      med=0s     max=31.97ms p(90)=0s      p(95)=0s    
     http_req_duration..............: avg=2.59ms  min=55.39µs med=2.6ms  max=37.08ms p(90)=2.9ms   p(95)=3.14ms
       { expected_response:true }...: avg=2.67ms  min=83.12µs med=2.6ms  max=37.08ms p(90)=2.91ms  p(95)=3.16ms
     http_req_failed................: 3.08%   ✓ 35217        ✗ 1104829
     http_req_receiving.............: avg=10.31µs min=3.2µs   med=8.18µs max=5.81ms  p(90)=12.71µs p(95)=14.5µs
     http_req_sending...............: avg=5.64µs  min=1.69µs  med=4.46µs max=34.6ms  p(90)=6.68µs  p(95)=7.43µs
     http_req_tls_handshaking.......: avg=0s      min=0s      med=0s     max=0s      p(90)=0s      p(95)=0s    
     http_req_waiting...............: avg=2.58ms  min=42.59µs med=2.59ms max=21.72ms p(90)=2.88ms  p(95)=3.12ms
     http_reqs......................: 1140046 37996.734938/s
     iteration_duration.............: avg=2.62ms  min=68.5µs  med=2.62ms max=40.52ms p(90)=2.93ms  p(95)=3.18ms
     iterations.....................: 1140046 37996.734938/s
     vus............................: 100     min=100        max=100  
     vus_max........................: 100     min=100        max=100  


running (0m30.0s), 000/100 VUs, 1140046 complete and 0 interrupted iterations
default ✓ [======================================] 100 VUs  30s

After:

$ k6 run script.js

          /\      |‾‾| /‾‾/   /‾‾/   
     /\  /  \     |  |/  /   /  /    
    /  \/    \    |     (   /   ‾‾\  
   /          \   |  |\  \ |  (‾)  | 
  / __________ \  |__| \__\ \_____/ .io

     execution: local
        script: script.js
        output: -

     scenarios: (100.00%) 1 scenario, 100 max VUs, 1m0s max duration (incl. graceful stop):
              * default: 100 looping VUs for 30s (gracefulStop: 30s)


     data_received..................: 180 MB  6.0 MB/s
     data_sent......................: 208 MB  6.9 MB/s
     http_req_blocked...............: avg=2.25µs min=250ns   med=601ns  max=48.18ms p(90)=952ns  p(95)=1.14µs 
     http_req_connecting............: avg=522ns  min=0s      med=0s     max=43.31ms p(90)=0s     p(95)=0s     
     http_req_duration..............: avg=2.06ms min=58.63µs med=1.88ms max=99.62ms p(90)=3.35ms p(95)=3.71ms 
       { expected_response:true }...: avg=2.06ms min=58.63µs med=1.88ms max=99.62ms p(90)=3.35ms p(95)=3.71ms 
     http_req_failed................: 0.00%   ✓ 0            ✗ 1431309
     http_req_receiving.............: avg=8.6µs  min=2.91µs  med=6.28µs max=6.4ms   p(90)=9.27µs p(95)=11.15µs
     http_req_sending...............: avg=4.84µs min=1.51µs  med=3.53µs max=44.83ms p(90)=5.32µs p(95)=6.18µs 
     http_req_tls_handshaking.......: avg=0s     min=0s      med=0s     max=0s      p(90)=0s     p(95)=0s     
     http_req_waiting...............: avg=2.05ms min=46.14µs med=1.87ms max=99.6ms  p(90)=3.33ms p(95)=3.7ms  
     http_reqs......................: 1431309 47707.976513/s
     iteration_duration.............: avg=2.09ms min=81.63µs med=1.9ms  max=99.65ms p(90)=3.37ms p(95)=3.73ms 
     iterations.....................: 1431309 47707.976513/s
     vus............................: 100     min=100        max=100  
     vus_max........................: 100     min=100        max=100  


running (0m30.0s), 000/100 VUs, 1431309 complete and 0 interrupted iterations
default ✓ [======================================] 100 VUs  30s

leonardo-albertovich · 2024-04-29T13:43:21Z

You are correct about the second request causing an issue but in the case of HTTP/1.x the stream instance is supposed to live through the entire session so I think these patches would be a better fit :

diff --git a/src/http_server/flb_http_server_http1.c b/src/http_server/flb_http_server_http1.c
index f5e2d735d..e6635f06a 100644
--- a/src/http_server/flb_http_server_http1.c
+++ b/src/http_server/flb_http_server_http1.c
@@ -138,6 +138,14 @@ static int http1_session_process_request(struct flb_http1_server_session *sessio
     struct mk_http_header  *header;
     int                     result;

+    result = flb_http_request_init(&session->stream.request);
+
+    if (result != 0) {
+        return -1;
+    }
+
+    session->stream.request.stream = &session->stream;
+
     if (session->inner_request.uri_processed.data != NULL) {
         session->stream.request.path = \
             cfl_sds_create_len(session->inner_request.uri_processed.data,
@@ -471,6 +479,8 @@ void flb_http1_server_session_destroy(struct flb_http1_server_session *session)
             session->inner_session.channel = NULL;
         }

+        flb_http_stream_destroy(&session->stream);
+
         session->initialized = FLB_FALSE;
     }
 }

diff --git a/src/flb_http_common.c b/src/flb_http_common.c
index fbd427335..6b629a6ec 100644
--- a/src/flb_http_common.c
+++ b/src/flb_http_common.c
@@ -437,6 +437,11 @@ void flb_http_stream_destroy(struct flb_http_stream *stream)
             cfl_list_del(&stream->_head);
         }

-        flb_free(stream);
+        flb_http_request_destroy(&stream->request);
+        flb_http_response_destroy(&stream->response);
+
+        if (stream->releasable) {
+            flb_free(stream);
+        }
     }
 }

This ensures that the request instance is properly initialized at all times while also removing a leak that could be triggered by prematurely closed connections.

@leonardo-albertovich

This patch addresses feedback from @leonardo-albertovich, reducing how many components are being recreated on each new request for a single session. Now it only recreates the request object, rather than creating a new stream for each request of a session. It replaces my original fix in 0ef074a Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

iandrewt · 2024-04-29T14:46:46Z

Makes sense, and works on my end too. Updated the PR with your patch!

Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

iandrewt added 5 commits April 25, 2024 20:37

in_http: enable keep-alive

1237eff

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_http plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

in_elasticsearch: enable keep-alive

61b5928

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_elasticsearch plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

in_opentelemetry: enable keep-alive

2890b8c

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_opentelemetry plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

in_splunk: enable keep-alive

3a2f91f

This patch stops the forced connection-close for requests that aren't HTTP/2 on the in_splunk plugin Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

github-actions bot added the docs-required label Apr 25, 2024

iandrewt marked this pull request as ready for review April 25, 2024 11:21

iandrewt requested review from edsiper, leonardo-albertovich, fujimotos and koleini as code owners April 25, 2024 11:21

iandrewt temporarily deployed to pr April 25, 2024 15:51 — with GitHub Actions Inactive

iandrewt temporarily deployed to pr April 25, 2024 16:12 — with GitHub Actions Inactive

iandrewt temporarily deployed to pr April 25, 2024 16:13 — with GitHub Actions Inactive

http_server: fix formatting of if/else block

4020f68

Signed-off-by: Andrew Titmuss <iandrewt@icloud.com>

leonardo-albertovich approved these changes Apr 29, 2024

View reviewed changes

iandrewt temporarily deployed to pr April 29, 2024 15:36 — with GitHub Actions Inactive

iandrewt temporarily deployed to pr April 29, 2024 15:58 — with GitHub Actions Inactive

iandrewt temporarily deployed to pr April 29, 2024 16:06 — with GitHub Actions Inactive

edsiper merged commit b2f2c93 into fluent:master Apr 30, 2024
44 checks passed

BrewTestBot mentioned this pull request May 21, 2024

fluent-bit 3.0.4 Homebrew/homebrew-core#172252

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

http_server: Fix keep-alive for HTTP/1.1 requests #8761

http_server: Fix keep-alive for HTTP/1.1 requests #8761

iandrewt commented Apr 25, 2024 •

edited

iandrewt commented Apr 25, 2024

iandrewt commented Apr 25, 2024

iandrewt commented Apr 25, 2024

iandrewt commented Apr 25, 2024

iandrewt commented Apr 26, 2024

leonardo-albertovich commented Apr 29, 2024

iandrewt commented Apr 29, 2024

http_server: Fix keep-alive for HTTP/1.1 requests #8761

http_server: Fix keep-alive for HTTP/1.1 requests #8761

Conversation

iandrewt commented Apr 25, 2024 • edited

iandrewt commented Apr 25, 2024

iandrewt commented Apr 25, 2024

iandrewt commented Apr 25, 2024

iandrewt commented Apr 25, 2024

iandrewt commented Apr 26, 2024

leonardo-albertovich commented Apr 29, 2024

iandrewt commented Apr 29, 2024

iandrewt commented Apr 25, 2024 •

edited