Skywalking installation and use

SkyWalking is an open source observable platform for collecting, analyzing, aggregating and visualizing data from services and cloud native infrastructure. SkyWalking provides an easy way to maintain a clear view of distributed systems, even across the cloud. It is a modern APM designed for cloud native, container based distributed systems.

Why?

SkyWalking provides a solution to observe and monitor distributed systems in many different scenarios. First, like the traditional way, SkyWalking is Java, c#, node JS, Go, PHP, Nginx LUA and other services provide automatic meter agents. (including calls for Python and C++ SDK contributions). In a multilingual and continuously deployed environment, the cloud native infrastructure becomes more powerful, but also more complex. SkyWalking's service grid receiver allows SkyWalking to receive telemetry data from service grid frameworks such as Istio/Envoy and Linkerd, so that users can understand the entire distributed system.

Download decompression

Official website: https://skywalking.apache.org/
Download link: https://skywalking.apache.org/downloads/

Because this is connected to elasticsearch7 x. So download apache-skywalking-apm-es7-8.5 tar. GZ installation package

decompression

/data/aimm/tools
tar -zxvf apache-skywalking-apm-es7-8.5.tar.gz

oap configuration

Configure the application. Here, 'zookeeper', 'consumer', 'etcd', 'nacos', etc. are supported

vim apache-skywalking-apm-bin-es7/config/application.yml

  elasticsearch7:
    nameSpace: ${SW_NAMESPACE:"skywalking"}
    clusterNodes: ${SW_STORAGE_ES_CLUSTER_NODES:172.16.253.10:9200} #The address of es is configured here
    protocol: ${SW_STORAGE_ES_HTTP_PROTOCOL:"http"}
    trustStorePath: ${SW_STORAGE_ES_SSL_JKS_PATH:""}
    trustStorePass: ${SW_STORAGE_ES_SSL_JKS_PASS:""}
    dayStep: ${SW_STORAGE_DAY_STEP:1} 
    indexShardsNumber: ${SW_STORAGE_ES_INDEX_SHARDS_NUMBER:1}
    indexReplicasNumber: ${SW_STORAGE_ES_INDEX_REPLICAS_NUMBER:1}
    superDatasetDayStep: ${SW_SUPERDATASET_STORAGE_DAY_STEP:-1}
    superDatasetIndexShardsFactor: ${SW_STORAGE_ES_SUPER_DATASET_INDEX_SHARDS_FACTOR:5}
    superDatasetIndexReplicasNumber: ${SW_STORAGE_ES_SUPER_DATASET_INDEX_REPLICAS_NUMBER:0}
    user: ${SW_ES_USER:"aimm_skywalking"}  #User name for configuration
    password: ${SW_ES_PASSWORD:"Pass!234"} #Configure the password for es
    secretsManagementFile: ${SW_ES_SECRETS_MANAGEMENT_FILE:""}
    bulkActions: ${SW_STORAGE_ES_BULK_ACTIONS:1000}
    flushInterval: ${SW_STORAGE_ES_FLUSH_INTERVAL:10}

For performance considerations, the official recommendation is in elasticsearch Add the following configuration to the YML configuration

thread_pool.index.queue_size: 1000 #ElasticSearch 6 only
thread_pool.write.queue_size: 1000 #For ElasticSearch 6 and 7

index.max_result_window: 1000000 #Remember to set when there is an error on the trace page

webapp configuration

The webapp configuration file is in / webapp / webapp In YML

vim apache-skywalking-apm-bin-es7/webapp/webapp.yml

server:
  port: 8080  #Port used to access the web

collector:
  path: /graphql
  ribbon:
    ReadTimeout: 10000
    # Point to all backend's restHost:restPort, split by ,
    listOfServers: 127.0.0.1:12800

By default, I use graphql to access the data collection port of oap, so I listen to port 12800. Because I deploy oap and webapp on the same server, the default address is 127.0 0.1. If it is deployed on different servers, it needs to be modified to the corresponding server address.

start-up

There are complete scripts in the / bin directory, which can be started through startup SH starts oap and webapp processes at the same time. What the script actually does is to call oapservice in the same directory SH and webappservice SH script. The script content is as follows. Therefore, if we consider deploying the two processes separately, we can call the corresponding script to run it separately.

vim apache-skywalking-apm-bin-es7/bin/startup.sh

PRG="$0"
PRGDIR=`dirname "$PRG"`
OAP_EXE=oapService.sh
WEBAPP_EXE=webappService.sh

"$PRGDIR"/"$OAP_EXE"

"$PRGDIR"/"$WEBAPP_EXE"

agent configuration

To use the agent, copy the entire directory of / agent to the corresponding server to be monitored, and modify the agent.config under / agent/config Config configuration

vim apache-skywalking-apm-bin-es7/agent/config/agent.config


# Different namespace s will cause the call link tracing to be interrupted
agent.namespace=${SW_AGENT_NAMESPACE:hmall}

# The name of the service displayed on the page can also be - dskywalking agent. service_ Name = XXX specified
agent.service_name=${SW_AGENT_NAME:gateway}

# The calling address of the platform can also be through - dskywalking collector. backend_ service=127.0. 0.1:80 specified
collector.backend_service=${SW_AGENT_COLLECTOR_BACKEND_SERVICES:172.16.0.11:11800}

# Ignore request collection for specified suffix
agent.ignore_suffix=${SW_AGENT_IGNORE_SUFFIX:.jpg,.jpeg,.js,.css,.png,.bmp,.gif,.ico,.mp3,.mp4,.html,.svg}


# Sampling rate every 3 seconds, negative number represents 100%
agent.sample_n_per_3_secs=${SW_AGENT_SAMPLE:-1}

After modification, add - javaagent: / data / AIMM / tools / agent / skywalking agent when starting the java process jar=agent. service_ Name = ${project_name} to load the agent, where ${project_name} represents the corresponding project name.

For example, a complete startup command

java -jar -Xmx512m -Xms512m -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/data/aimm/log/aimm-operation-trace-app/ -Djava.io.tmpdir=/data/aimm/tmp -javaagent:/data/aimm/tools/agent/skywalking-agent.jar=agent.service_name=aimm-operation-trace-app -Dspring.profiles.active=test /data/aimm/app/aimm-operation-trace-app-2.2.5-SNAPSHOT.jar

SkyWalking is logically divided into four parts: probe, platform backend, storage and UI.

![kubernetes](https://cdn.jsdelivr.net/gh/heyonggs/images@master/wp/2021/skywalking/skywalking-1.jpg)

Probes: probes collect data and reformat them according to SkyWalking requirements (different probes support different sources).
Platform back end: the platform back end supports data aggregation, analysis and stream processing, including tracking, indicators and logs.
Storage: store SkyWalking data through an open / pluggable interface. You can choose an existing implementation, such as ElasticSearch, H2, MySQL, TiDB, InfluxDB, or implement your own. Welcome to the patch for the new storage implementer!
UI: UI is a highly customizable Web-based interface that allows SkyWalking end users to visualize and manage SkyWalking data.

Keywords: Java skywalking

Added by Saruman on Thu, 30 Dec 2021 05:30:50 +0200

Programming VIP