kazusa
/
skywalking-java
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
skywalking-java/docs/en/setup/backend/backend-storage.md

15 KiB
Raw Blame History

Backend storage

SkyWalking storage is pluggable, we have provided the following storage solutions, you could easily use is by changing the application.yml

Native supported storage

  • H2
  • ElasticSearch 6, 7
  • MySQL
  • TiDB
  • InfluxDB

Redistribution version with supported storage.

  • ElasticSearch 5

H2

Active H2 as storage, set storage provider to H2 In-Memory Databases. Default in distribution package. Please read Database URL Overview in H2 official document, you could set the target to H2 in Embedded, Server and Mixed modes.

Setting fragment example

storage:
  h2:
    driver: org.h2.jdbcx.JdbcDataSource
    url: jdbc:h2:mem:skywalking-oap-db
    user: sa

ElasticSearch

  • In order to activate ElasticSearch 6 as storage, set storage provider to elasticsearch
  • In order to activate ElasticSearch 7 as storage, set storage provider to elasticsearch7

Required ElasticSearch 6.3.2 or higher. HTTP RestHighLevelClient is used to connect server.

  • For ElasticSearch 6.3.2 ~ 7.0.0 (excluded), please download the apache-skywalking-bin.tar.gz or apache-skywalking-bin.zip,
  • For ElasticSearch 7.0.0 ~ 8.0.0 (excluded), please download the apache-skywalking-bin-es7.tar.gz or apache-skywalking-bin-es7.zip.

For now, ElasticSearch 6 and ElasticSearch 7 share the same configurations, as follows:

Setting fragment example

storage:
  elasticsearch:
    # nameSpace: ${SW_NAMESPACE:""}
    # user: ${SW_ES_USER:""} # User needs to be set when Http Basic authentication is enabled
    # password: ${SW_ES_PASSWORD:""} # Password to be set when Http Basic authentication is enabled
    #trustStorePath: ${SW_SW_STORAGE_ES_SSL_JKS_PATH:""}
    #trustStorePass: ${SW_SW_STORAGE_ES_SSL_JKS_PASS:""}
    enablePackedDownsampling: ${SW_STORAGE_ENABLE_PACKED_DOWNSAMPLING:true} # Hour and Day metrics will be merged into minute index.
    dayStep: ${SW_STORAGE_DAY_STEP:1} # Represent the number of days in the one minute/hour/day index.
    clusterNodes: ${SW_STORAGE_ES_CLUSTER_NODES:localhost:9200}
    protocol: ${SW_STORAGE_ES_HTTP_PROTOCOL:"http"}
    indexShardsNumber: ${SW_STORAGE_ES_INDEX_SHARDS_NUMBER:2}
    indexReplicasNumber: ${SW_STORAGE_ES_INDEX_REPLICAS_NUMBER:0}
    # Those data TTL settings will override the same settings in core module.
    recordDataTTL: ${SW_STORAGE_ES_RECORD_DATA_TTL:7} # Unit is day
    otherMetricsDataTTL: ${SW_STORAGE_ES_OTHER_METRIC_DATA_TTL:45} # Unit is day
    monthMetricsDataTTL: ${SW_STORAGE_ES_MONTH_METRIC_DATA_TTL:18} # Unit is month
    # Batch process setting, refer to https://www.elastic.co/guide/en/elasticsearch/client/java-api/5.5/java-docs-bulk-processor.html
    bulkActions: ${SW_STORAGE_ES_BULK_ACTIONS:2000} # Execute the bulk every 2000 requests
    bulkSize: ${SW_STORAGE_ES_BULK_SIZE:20} # flush the bulk every 20mb
    flushInterval: ${SW_STORAGE_ES_FLUSH_INTERVAL:10} # flush the bulk every 10 seconds whatever the number of requests
    concurrentRequests: ${SW_STORAGE_ES_CONCURRENT_REQUESTS:2} # the number of concurrent requests
    resultWindowMaxSize: ${SW_STORAGE_ES_QUERY_MAX_WINDOW_SIZE:10000}
    metadataQueryMaxSize: ${SW_STORAGE_ES_QUERY_MAX_SIZE:5000}
    segmentQueryMaxSize: ${SW_STORAGE_ES_QUERY_SEGMENT_SIZE:200}
    profileTaskQueryMaxSize: ${SW_STORAGE_ES_QUERY_PROFILE_TASK_SIZE:200}
    advanced: ${SW_STORAGE_ES_ADVANCED:""}

Downsampling Data Packing

Downsampling data packing(storage/elasticsearch/enablePackedDownsampling, default activated) is a new feature since 7.0.0.
Metrics data has 4 different precisions,based on core/default/downsampling configurations. In previous(6.x), every precision of each metrics had one separated index. After this is activated, metrics of day and hour precisions are merged into minute precision. The number of indexes decreased, and cause less payload to the ElasticSearch server.

ElasticSearch 6 With Https SSL Encrypting communications.

example:

storage:
  elasticsearch:
    # nameSpace: ${SW_NAMESPACE:""}
    user: ${SW_ES_USER:""} # User needs to be set when Http Basic authentication is enabled
    password: ${SW_ES_PASSWORD:""} # Password to be set when Http Basic authentication is enabled
    clusterNodes: ${SW_STORAGE_ES_CLUSTER_NODES:localhost:443}
    trustStorePath: ${SW_SW_STORAGE_ES_SSL_JKS_PATH:"../es_keystore.jks"}
    trustStorePass: ${SW_SW_STORAGE_ES_SSL_JKS_PASS:""}
    protocol: ${SW_STORAGE_ES_HTTP_PROTOCOL:"https"}
    indexShardsNumber: ${SW_STORAGE_ES_INDEX_SHARDS_NUMBER:2}
    indexReplicasNumber: ${SW_STORAGE_ES_INDEX_REPLICAS_NUMBER:0}
    # Those data TTL settings will override the same settings in core module.
    recordDataTTL: ${SW_STORAGE_ES_RECORD_DATA_TTL:7} # Unit is day
    otherMetricsDataTTL: ${SW_STORAGE_ES_OTHER_METRIC_DATA_TTL:45} # Unit is day
    monthMetricsDataTTL: ${SW_STORAGE_ES_MONTH_METRIC_DATA_TTL:18} # Unit is month
    # Batch process setting, refer to https://www.elastic.co/guide/en/elasticsearch/client/java-api/5.5/java-docs-bulk-processor.html
    bulkActions: ${SW_STORAGE_ES_BULK_ACTIONS:2000} # Execute the bulk every 2000 requests
    bulkSize: ${SW_STORAGE_ES_BULK_SIZE:20} # flush the bulk every 20mb
    flushInterval: ${SW_STORAGE_ES_FLUSH_INTERVAL:10} # flush the bulk every 10 seconds whatever the number of requests
    concurrentRequests: ${SW_STORAGE_ES_CONCURRENT_REQUESTS:2} # the number of concurrent requests
    advanced: ${SW_STORAGE_ES_ADVANCED:""}

Data TTL

TTL in ElasticSearch overrides the settings of core, read ElasticSearch section in TTL document

Daily Index Step

Daily index step(storage/elasticsearch/dayStep, default 1) represents the index creation period. In this period, several days(dayStep value)' metrics are saved.

Mostly, users don't need to change the value manually. As SkyWalking is designed to observe large scale distributed system. But in some specific cases, users want to set a long TTL value, such as more than 60 days, but their ElasticSearch cluster isn't powerful due to the low traffic in the production environment. This value could be increased to 5(or more), if users could make sure single one index could support these days(5 in this case) metrics and traces.

Such as, if dayStep == 11,

  1. data in [2000-01-01, 2000-01-11] will be merged into the index-20000101.
  2. data in [2000-01-12, 2000-01-22] will be merged into the index-20000112.

NOTICE, TTL deletion would be affected by these. You should set an extra more dayStep in your TTL. Such as you want to TTL == 30 days and dayStep == 10, you actually need to set TTL = 40;

Advanced Configurations For Elasticsearch Index

You can add advanced configurations in JSON format to set ElasticSearch index settings by following ElasticSearch doc

For example, set translog settings:

storage:
  elasticsearch:
    # ......
    advanced: ${SW_STORAGE_ES_ADVANCED:"{\"index.translog.durability\":\"request\",\"index.translog.sync_interval\":\"5s\"}"}

Recommended ElasticSearch server-side configurations

You could add following config to elasticsearch.yml, set the value based on your env.

# In tracing scenario, consider to set more than this at least.
thread_pool.index.queue_size: 1000 # Only suitable for ElasticSearch 6
thread_pool.write.queue_size: 1000 # Suitable for ElasticSearch 6 and 7

# When you face query error at trace page, remember to check this.
index.max_result_window: 1000000

We strongly advice you to read more about these configurations from ElasticSearch official document. This effects the performance of ElasticSearch very much.

ElasticSearch 6 with Zipkin trace extension

This implementation shares most of elasticsearch, just extend to support zipkin span storage. It has all same configs.

storage:
  zipkin-elasticsearch:
    nameSpace: ${SW_NAMESPACE:""}
    clusterNodes: ${SW_STORAGE_ES_CLUSTER_NODES:localhost:9200}
    protocol: ${SW_STORAGE_ES_HTTP_PROTOCOL:"http"}
    user: ${SW_ES_USER:""}
    password: ${SW_ES_PASSWORD:""}
    indexShardsNumber: ${SW_STORAGE_ES_INDEX_SHARDS_NUMBER:2}
    indexReplicasNumber: ${SW_STORAGE_ES_INDEX_REPLICAS_NUMBER:0}
    # Those data TTL settings will override the same settings in core module.
    recordDataTTL: ${SW_STORAGE_ES_RECORD_DATA_TTL:7} # Unit is day
    otherMetricsDataTTL: ${SW_STORAGE_ES_OTHER_METRIC_DATA_TTL:45} # Unit is day
    monthMetricsDataTTL: ${SW_STORAGE_ES_MONTH_METRIC_DATA_TTL:18} # Unit is month
    # Batch process setting, refer to https://www.elastic.co/guide/en/elasticsearch/client/java-api/5.5/java-docs-bulk-processor.html
    bulkActions: ${SW_STORAGE_ES_BULK_ACTIONS:2000} # Execute the bulk every 2000 requests
    bulkSize: ${SW_STORAGE_ES_BULK_SIZE:20} # flush the bulk every 20mb
    flushInterval: ${SW_STORAGE_ES_FLUSH_INTERVAL:10} # flush the bulk every 10 seconds whatever the number of requests
    concurrentRequests: ${SW_STORAGE_ES_CONCURRENT_REQUESTS:2} # the number of concurrent requests

ElasticSearch 6 with Jaeger trace extension

This implementation shares most of elasticsearch, just extend to support zipkin span storage. It has all same configs.

storage:
  jaeger-elasticsearch:
    nameSpace: ${SW_NAMESPACE:""}
    clusterNodes: ${SW_STORAGE_ES_CLUSTER_NODES:localhost:9200}
    protocol: ${SW_STORAGE_ES_HTTP_PROTOCOL:"http"}
    user: ${SW_ES_USER:""}
    password: ${SW_ES_PASSWORD:""}
    indexShardsNumber: ${SW_STORAGE_ES_INDEX_SHARDS_NUMBER:2}
    indexReplicasNumber: ${SW_STORAGE_ES_INDEX_REPLICAS_NUMBER:0}
    # Those data TTL settings will override the same settings in core module.
    recordDataTTL: ${SW_STORAGE_ES_RECORD_DATA_TTL:7} # Unit is day
    otherMetricsDataTTL: ${SW_STORAGE_ES_OTHER_METRIC_DATA_TTL:45} # Unit is day
    monthMetricsDataTTL: ${SW_STORAGE_ES_MONTH_METRIC_DATA_TTL:18} # Unit is month
    # Batch process setting, refer to https://www.elastic.co/guide/en/elasticsearch/client/java-api/5.5/java-docs-bulk-processor.html
    bulkActions: ${SW_STORAGE_ES_BULK_ACTIONS:2000} # Execute the bulk every 2000 requests
    bulkSize: ${SW_STORAGE_ES_BULK_SIZE:20} # flush the bulk every 20mb
    flushInterval: ${SW_STORAGE_ES_FLUSH_INTERVAL:10} # flush the bulk every 10 seconds whatever the number of requests
    concurrentRequests: ${SW_STORAGE_ES_CONCURRENT_REQUESTS:2} # the number of concurrent requests

About Namespace

When namespace is set, names of all indexes in ElasticSearch will use it as prefix.

About Authentication

We only support basic authentication. If you need that, you could set user and password. For how to enable http basic authentication, you could read this https://brudtkuhl.com/blog/securing-elasticsearch/

MySQL

Active MySQL as storage, set storage provider to mysql.

NOTICE: MySQL driver is NOT allowed in Apache official distribution and source codes. Please download MySQL driver by yourself. Copy the connection driver jar to oap-libs.

storage:
  mysql:
    properties:
      jdbcUrl: ${SW_JDBC_URL:"jdbc:mysql://localhost:3306/swtest"}
      dataSource.user: ${SW_DATA_SOURCE_USER:root}
      dataSource.password: ${SW_DATA_SOURCE_PASSWORD:root@1234}
      dataSource.cachePrepStmts: ${SW_DATA_SOURCE_CACHE_PREP_STMTS:true}
      dataSource.prepStmtCacheSize: ${SW_DATA_SOURCE_PREP_STMT_CACHE_SQL_SIZE:250}
      dataSource.prepStmtCacheSqlLimit: ${SW_DATA_SOURCE_PREP_STMT_CACHE_SQL_LIMIT:2048}
      dataSource.useServerPrepStmts: ${SW_DATA_SOURCE_USE_SERVER_PREP_STMTS:true}
    metadataQueryMaxSize: ${SW_STORAGE_MYSQL_QUERY_MAX_SIZE:5000}

All connection related settings including link url, username and password are in application.yml. Here are some of the settings, please follow HikariCP connection pool document for all the settings.

TiDB

Currently tested TiDB in version 2.0.9, and Mysql Client driver in version 8.0.13. Active TiDB as storage, set storage provider to mysql.

storage:
  mysql:
    properties:
      jdbcUrl: ${SW_JDBC_URL:"jdbc:mysql://localhost:3306/swtest"}
      dataSource.user: ${SW_DATA_SOURCE_USER:root}
      dataSource.password: ${SW_DATA_SOURCE_PASSWORD:root@1234}
      dataSource.cachePrepStmts: ${SW_DATA_SOURCE_CACHE_PREP_STMTS:true}
      dataSource.prepStmtCacheSize: ${SW_DATA_SOURCE_PREP_STMT_CACHE_SQL_SIZE:250}
      dataSource.prepStmtCacheSqlLimit: ${SW_DATA_SOURCE_PREP_STMT_CACHE_SQL_LIMIT:2048}
      dataSource.useServerPrepStmts: ${SW_DATA_SOURCE_USE_SERVER_PREP_STMTS:true}
    metadataQueryMaxSize: ${SW_STORAGE_MYSQL_QUERY_MAX_SIZE:5000}

All connection related settings including link url, username and password are in application.yml. These settings can refer to the configuration of MySQL above.

InfluxDB

InfluxDB as storage since SkyWalking 7.0. It depends on H2/MySQL storage-plugin to store metadata like Inventory and ProfileTask. So, when we set InfluxDB as storage provider. We need to configure properties of InfluxDB and Metabase.

storage:
  influx:
    # Metadata storage provider configuration
    metabaseType: ${SW_STORAGE_METABASE_TYPE:H2} # There are 2 options as Metabase provider, H2 or MySQL.
    h2Props:
      dataSourceClassName: ${SW_STORAGE_METABASE_DRIVER:org.h2.jdbcx.JdbcDataSource}
      dataSource.url: ${SW_STORAGE_METABASE_URL:jdbc:h2:mem:skywalking-oap-db}
      dataSource.user: ${SW_STORAGE_METABASE_USER:sa}
      dataSource.password: ${SW_STORAGE_METABASE_PASSWORD:}
    mysqlProps:
      jdbcUrl: ${SW_STORAGE_METABASE_URL:"jdbc:mysql://localhost:3306/swtest"}
      dataSource.user: ${SW_STORAGE_METABASE_USER:root}
      dataSource.password: ${SW_STORAGE_METABASE_PASSWORD:root@1234}
      dataSource.cachePrepStmts: ${SW_STORAGE_METABASE_CACHE_PREP_STMTS:true}
      dataSource.prepStmtCacheSize: ${SW_STORAGE_METABASE_PREP_STMT_CACHE_SQL_SIZE:250}
      dataSource.prepStmtCacheSqlLimit: ${SW_STORAGE_METABASE_PREP_STMT_CACHE_SQL_LIMIT:2048}
      dataSource.useServerPrepStmts: ${SW_STORAGE_METABASE_USE_SERVER_PREP_STMTS:true}
    metadataQueryMaxSize: ${SW_STORAGE_METABASE_QUERY_MAX_SIZE:5000}
    # InfluxDB configuration
    url: ${SW_STORAGE_INFLUXDB_URL:http://localhost:8086}
    user: ${SW_STORAGE_INFLUXDB_USER:root}
    password: ${SW_STORAGE_INFLUXDB_PASSWORD:}
    database: ${SW_STORAGE_INFLUXDB_DATABASE:skywalking}
    actions: ${SW_STORAGE_INFLUXDB_ACTIONS:1000} # the number of actions to collect
    duration: ${SW_STORAGE_INFLUXDB_DURATION:1000} # the time to wait at most (milliseconds)
    fetchTaskLogMaxSize: ${SW_STORAGE_INFLUXDB_FETCH_TASK_LOG_MAX_SIZE:5000} # the max number of fetch task log in a request

All connection related settings including link url, username and password are in application.yml. The Metadata storage provider settings can refer to the configuration of H2/MySQL above.

ElasticSearch 5

ElasticSearch 5 is incompatible with ElasticSearch 6 Java client jar, so it could not be included in native distribution. OpenSkyWalking/SkyWalking-With-Es5x-Storage repo includes the distribution version.

More storage solution extension

Follow Storage extension development guide in Project Extensions document in development guide.