千家信息网

使用 gor 复制真实流量压测ES集群

发表于:2025-01-31 作者:千家信息网编辑
千家信息网最后更新 2025年01月31日,生产环境下, 我们一般使用 esrally来做es的基准测试。 但是毕竟和真实生产的请求场景可能有差异的。因此,有必要考虑使用真实流量来压测ES集群。常用的流量复制工具有tcpcopy之类的,这里我们
千家信息网最后更新 2025年01月31日使用 gor 复制真实流量压测ES集群

生产环境下, 我们一般使用 esrally来做es的基准测试。 但是毕竟和真实生产的请求场景可能有差异的。

因此,有必要考虑使用真实流量来压测ES集群。


常用的流量复制工具有tcpcopy之类的,这里我们介绍一个更简单易用的工具 gor 。

github地址: https://github.com/buger/goreplay


TIPS:

我在测试环境的 192.168.2.4 上实际测试用的命令(放大1倍流量):

# 192.168.2.4 为老es集群节点# 192.168.2.188 为新es集群节点wget https://github.com/buger/goreplay/releases/download/v1.0.0/gor_1.0.0_x64.tar.gztar xf gor_1.0.0_x64.tar.gz -C ././gor --input-raw :9200 --output-http "http://192.168.2.188:9200|200%"


下图是gor 输出到stdout的内容:


下图是我在老es集群更新文档后,可以看到新的es集群也自动更新了文档



更多的关于gor 流量复制的介绍,可以看下面的内容:


参考文档:

https://testerhome.com/articles/21180

https://github.com/buger/goreplay/wiki

https://www.jianshu.com/p/57e058ad4995




1、流量实时复制引流(--input-raw 拦截端口配合--output-http输出),例如将本机9200端口的HTTP流量实时复制到 es-dest:9200。

gor --input-raw :9200 --output-http "http://es-dest:9200"


2、如果目标服务器使用的库与线上机器一样,且只需要引流Get方法的请求。

gor --input-raw :9200  --http-allow-method GET --output-http "http://es-dest:9200"


3、当需要对线上服务进行整体性能压测时,可将线上请求扩大N倍,进行引流,将请求扩大1倍,也可缩小,调整"|"后面的百分比即可。

gor --input-raw :9200  --http-allow-method GET --output-http "http://es-dest:9200|200%"


4、只复制某个URL请求,--http-allow-url参数,-http-allow-url参数可用正则表达式(--output-http-url-regexp在gor 0.16已经过期,使用--http-allow-url代替)。

gor --input-raw :9200 --http-allow-method GET --output-http "http://es-dest:9200" --http-allow-url mall.*hotword


5、多目标服务器的流量复制引流,有点类似nginx的mirror。

gor --input-raw :9200 --output-http "http://es-dest:9200" --output-http "http://es-dest:9200"


6、将流量转发到多个端点,默认情况下,它会将相同的流量发送到所有输出。

gor --input-tcp :2920020 --output-http "http://staging.com"  --output-http "http://dev.com"


7、将相同的流量发送到多个站点,并且平分所有流量。

gor --input-raw :9200 --output-http "http://staging.com"  --output-http "http://dev.com" --split-output true


8、HTTP超时,默认情况下,http请求和响应的超时时间为5秒。你可以像这样覆盖它:

gor --input-tcp replay.local:2920020 --output-http http://staging.com --output-http-timeout 30s


9、基本身份验证,如果您的开发或登台环境受基本身份验证保护,那么可以在重放期间注入这些凭据:

gor --input-raw :9200 --output-http "http://user:pass@staging.com"


10、性能压力测试

可以将流量复制到文件,然后再对他们进行回放。回放的时候,流量会维持原始的时间间隔。如果你使用了百分比来进行速率限制,那么回放的速率会相应的增加或减少。有了这种速率限制,gor就可以用来进行压力测试。

gor --input-file "requests.gor|200%" --output-http "staging.com"

目前,input-file仅在使用基于百分比的限制器时才支持此功能。与默认限制器不同input-file,它不会降低请求速度,而会减慢速度或加速请求发射。


11 、录制与回放

./gor --input-raw :9200 --output-file=requests.gor

执行录制命令后,将会创建新文件并不断向其写入所有捕获的请求。


12、流量回放,重播来自文件的请求。

./gor --input-file requests.gor --output-http="http://localhost:92000"


您应该看到所有记录到http://localhost:92000的请求,并且它们将以相同的顺序重播,并且与录制的时间完全相同。



gor的用法也远不止上述提到的这些,更多的应用测试场景,还请读者自行挖掘。






./gor --help 可以列出更全面的帮助信息

Gor is a simple http traffic replication tool written in Go. Its main goal is to replay traffic from production servers to staging and dev environments.Project page: https://github.com/buger/gorAuthor:  leonsbox@gmail.comCurrent Version: 1.0.0  -copy-buffer-size int    Set the buffer size for an individual request (default 5M) (default 524289200)  -cpuprofile string    write cpu profile to file  -debug verbose    Turn on debug output, shows all intercepted traffic. Works only when with verbose flag  -exit-after duration    exit after specified duration  -http-allow-header value    A regexp to match a specific header against. Requests with non-matching headers will be dropped:     gor --input-raw :9200 --output-http staging.com --http-allow-header api-version:^v1  -http-allow-method value    Whitelist of HTTP methods to replay. Anything else will be dropped:    gor --input-raw :9200 --output-http staging.com --http-allow-method GET --http-allow-method OPTIONS  -http-allow-url value    A regexp to match requests against. Filter get matched against full url with domain. Anything else will be dropped:     gor --input-raw :9200 --output-http staging.com --http-allow-url ^www.  -http-basic-auth-filter value    A regexp to match the decoded basic auth string against. Requests with non-matching headers will be dropped:     gor --input-raw :9200 --output-http staging.com --http-basic-auth-filter "^customer[0-9].*"  -http-disallow-header value    A regexp to match a specific header against. Requests with matching headers will be dropped:     gor --input-raw :9200 --output-http staging.com --http-disallow-header "User-Agent: Replayed by Gor"  -http-disallow-url value    A regexp to match requests against. Filter get matched against full url with domain. Anything else will be forwarded:     gor --input-raw :9200 --output-http staging.com --http-disallow-url ^www.  -http-header-limiter value    Takes a fraction of requests, consistently taking or rejecting a request based on the FNV32-1A hash of a specific header:     gor --input-raw :9200 --output-http staging.com --http-header-limiter user-id:25%  -http-original-host    Normally gor replaces the Host http header with the host supplied with --output-http.  This option disables that behavior, preserving the original Host header.  -http-param-limiter value    Takes a fraction of requests, consistently taking or rejecting a request based on the FNV32-1A hash of a specific GET param:     gor --input-raw :9200 --output-http staging.com --http-param-limiter user_id:25%  -http-pprof :8181    Enable profiling. Starts  http server on specified port, exposing special /debug/pprof endpoint. Example: :8181  -http-rewrite-header value    Rewrite the request header based on a mapping:    gor --input-raw :9200 --output-http staging.com --http-rewrite-header Host: (.*).example.com,$1.beta.example.com  -http-rewrite-url value    Rewrite the request url based on a mapping:    gor --input-raw :9200 --output-http staging.com --http-rewrite-url /v1/user/([^\/]+)/ping:/v2/user/$1/ping  -http-set-header value    Inject additional headers to http reqest:    gor --input-raw :9200 --output-http staging.com --http-set-header 'User-Agent: Gor'  -http-set-param value    Set request url param, if param already exists it will be overwritten:    gor --input-raw :9200 --output-http staging.com --http-set-param api_key=1  -input-dummy value    Used for testing outputs. Emits 'Get /' request every 1s  -input-file value    Read requests from file:     gor --input-file ./requests.gor --output-http staging.com  -input-file-loop    Loop input files, useful for performance testing.  -input-kafka-host string    Send request and response stats to Kafka:    gor --output-stdout --input-kafka-host '192.168.0.1:9092,192.168.0.2:9092'  -input-kafka-json-format    If turned on, it will assume that messages coming in JSON format rather than  GoReplay text format.  -input-kafka-topic string    Send request and response stats to Kafka:    gor --output-stdout --input-kafka-topic 'kafka-log'  -input-raw value    Capture traffic from given port (use RAW sockets and require *sudo* access):    # Capture traffic from 9200 port    gor --input-raw :9200 --output-http staging.com  -input-raw-bpf-filter string    BPF filter to write custom expressions. Can be useful in case of non standard network interfaces like tunneling or SPAN port. Example: --input-raw-bpf-filter 'dst port 9200'  -input-raw-buffer-size int    Controls size of the OS buffer (in bytes) which holds packets until they dispatched. Default value depends by system: in Linux around 2MB. If you see big package drop, increase this value.  -input-raw-engine libpcap    Intercept traffic using libpcap (default), and `raw_socket` (default "libpcap")  -input-raw-expire duration    How much it should wait for the last TCP packet, till consider that TCP message complete. (default 2s)  -input-raw-immediate-mode    Set pcap interface to immediate mode.  -input-raw-override-snaplen    Override the capture snaplen to be 64k. Required for some Virtualized environments  -input-raw-realip-header string    If not blank, injects header with given name and real IP value to the request payload. Usually this header should be named: X-Real-IP  -input-raw-timestamp-type string    Possible values: PCAP_TSTAMP_HOST, PCAP_TSTAMP_HOST_LOWPREC, PCAP_TSTAMP_HOST_HIPREC, PCAP_TSTAMP_ADAPTER, PCAP_TSTAMP_ADAPTER_UNSYNCED. This values not supported on all systems, GoReplay will tell you available values of you put wrong one.  -input-raw-track-response    If turned on Gor will track responses in addition to requests, and they will be available to middleware and file output.  -input-tcp value    Used for internal communication between Gor instances. Example:     # Receive requests from other Gor instances on 2920020 port, and redirect output to staging    gor --input-tcp :2920020 --output-http staging.com  -input-tcp-certificate string    Path to PEM encoded certificate file. Used when TLS turned on.  -input-tcp-certificate-key string    Path to PEM encoded certificate key file. Used when TLS turned on.  -input-tcp-secure    Turn on TLS security. Do not forget to specify certificate and key files.  -memprofile string    write memory profile to this file  -middleware string    Used for modifying traffic using external command  -output-dummy value    DEPRECATED: use --output-stdout instead  -output-file value    Write incoming requests to file:     gor --input-raw :9200 --output-file ./requests.gor  -output-file-append    The flushed chunk is appended to existence file or not.   -output-file-flush-interval duration    Interval for forcing buffer flush to the file, default: 1s. (default 1s)  -output-file-max-size-limit value    Max size of output file, Default: 1TB (default -1)  -output-file-queue-limit int    The length of the chunk queue. Default: 256 (default 256)  -output-file-size-limit value    Size of each chunk. Default: 32mb (default 33554432)  -output-http value    Forwards incoming requests to given http address.    # Redirect all incoming requests to staging.com address     gor --input-raw :9200 --output-http http://staging.com  -output-http-compatibility-mode    Use standard Go client, instead of built-in implementation. Can be slower, but more compatible.  -output-http-debug    Enables http debug output.  -output-http-elasticsearch string    Send request and response stats to ElasticSearch:    gor --input-raw :9200 --output-http staging.com --output-http-elasticsearch 'es_host:api_port/index_name'  -output-http-header --output-http-header    WARNING: --output-http-header DEPRECATED, use `--http-set-header` instead  -output-http-header-filter --output-http-header-filter    WARNING: --output-http-header-filter DEPRECATED, use `--http-allow-header` instead  -output-http-header-hash-filter output-http-header-hash-filter    WARNING: output-http-header-hash-filter DEPRECATED, use `--http-header-hash-limiter` instead  -output-http-method --output-http-method    WARNING: --output-http-method DEPRECATED, use `--http-allow-method` instead  -output-http-queue-len int    Number of requests that can be queued for output, if all workers are busy. default = 1000 (default 1000)  -output-http-redirects int    Enable how often redirects should be followed.  -output-http-response-buffer int    HTTP response buffer size, all data after this size will be discarded.  -output-http-rewrite-url --output-http-rewrite-url    WARNING: --output-http-rewrite-url DEPRECATED, use `--http-rewrite-url` instead  -output-http-stats    Report http output queue stats to console every N milliseconds. See output-http-stats-ms  -output-http-stats-ms int    Report http output queue stats to console every N milliseconds. default: 5000 (default 5000)  -output-http-timeout duration    Specify HTTP request/response timeout. By default 5s. Example: --output-http-timeout 30s (default 5s)  -output-http-track-response    If turned on, HTTP output responses will be set to all outputs like stdout, file and etc.  -output-http-url-regexp --output-http-url-regexp    WARNING: --output-http-url-regexp DEPRECATED, use `--http-allow-url` instead  -output-http-workers int    Gor uses dynamic worker scaling. Enter a number to set a maximum number of workers. default = 0 = unlimited.  -output-http-workers-min int    Gor uses dynamic worker scaling. Enter a number to set a minimum number of workers. default = 1.  -output-kafka-host string    Read request and response stats from Kafka:    gor --input-raw :9200 --output-kafka-host '192.168.0.1:9092,192.168.0.2:9092'  -output-kafka-json-format    If turned on, it will serialize messages from GoReplay text format to JSON.  -output-kafka-topic string    Read request and response stats from Kafka:    gor --input-raw :9200 --output-kafka-topic 'kafka-log'  -output-null    Used for testing inputs. Drops all requests.  -output-stdout    Used for testing inputs. Just prints to console data coming from inputs.  -output-tcp value    Used for internal communication between Gor instances. Example:     # Listen for requests on 9200 port and forward them to other Gor instance on 2920020 port    gor --input-raw :9200 --output-tcp replay.local:2920020  -output-tcp-secure    Use TLS secure connection. --input-file on another end should have TLS turned on as well.  -output-tcp-stats    Report TCP output queue stats to console every 5 seconds.  -prettify-http    If enabled, will automatically decode requests and responses with: Content-Encodning: gzip and Transfer-Encoding: chunked. Useful for debugging, in conjuction with --output-stdout  -split-output true    By default each output gets same traffic. If set to true it splits traffic equally among all outputs.  -stats    Turn on queue stats output  -verbose    Turn on more verbose output







0