Skip to content

ArpNetworking/metrics-aggregator-daemon

Repository files navigation

Metrics Aggregator Daemon

License: Apache 2 Jenkins Build Maven Artifact Docker

Aggregates samples into configurable time buckets (e.g. 1 second, 1 minute, etc.) published by metrics client libraries (e.g. Java, NodeJS, Ruby, etc.) to compute a variety of statistics. The statistics are reaggregatable and are published together with supporting data to configurable destination(s).

Usage

Installation

Manual

The artifacts from the build are in metrics-aggregator-daemon/target/appassembler and should be copied to an appropriate directory on your application host(s).

Docker

If you use Docker, we publish a base docker image that makes it easy for you to layer configuration on top of. Create a Docker image based on the image arpnetworking/mad. Configuration files are typically located at /opt/mad/config/ with pipeline files located at /opt/mad/config/pipelines. In addition, you can specify CONFIG_FILE (defaults to /opt/mad/config/config.conf), LOGGING_CONFIG (defaults to "-Dlogback.configurationFile=/opt/mad/config/logback.xml"), and JAVA_OPTS (defaults to "") environment variables to control startup.

Execution

In the installation's bin directory there are scripts to start Metrics Aggregator Daemon: mad (Linux) and mad.bat (Windows). One of these should be executed on system start with appropriate parameters; for example:

/usr/local/lib/metrics-aggregator-daemon/bin/mad /usr/local/lib/metrics-aggregator-daemon/config/config.conf

Configuration

Logging

To customize logging you may provide a LogBack configuration file. The project ships with logback.xml which writes logs to rotated files and with logback-console.xml which writes logs to STDOUT.

Outside of Docker, set the JAVA_OPTS environment variable to configure logging:

export JAVA_OPTS="-Dlogback.configurationFile=/usr/local/lib/metrics-aggregator-daemon/config/logback-console.xml"

Where /usr/local/lib/metrics-aggregator-daemon/config/logger-console.xml is the path to your logging configuration file.

Under Docker, set the LOGBACK_CONFIG environment variable to configure logging:

docker run -e LOGBACK_CONFIG=/opt/mad/config/logack-console.xml arpnetworking/mad:latest

Daemon

The Metrics Aggregator Daemon configuration is specified in a JSON file. The location of the configuration file is passed to mad as a command line argument:

/usr/local/lib/metrics-aggregator-daemon/config/config.json

The configuration specifies:

  • logDirectory - The location of additional logs. This is independent of the logging configuration.
  • pipelinesDirectory - The location of configuration files for each metrics pipeline.
  • httpHost - The ip address to bind the http server to.
  • httpPort - The port to bind the http server to.
  • httpHealthCheckPath - The path in the http server for the health check.
  • httpStatusPath - The path in the http server for the status.
  • jvmMetricsCollectionInterval - The JVM metrics collection interval in ISO-8601 period notation.
  • limiters - Configuration of zero or more limiters by name.
  • pekkoConfiguration - Configuration of Pekko.

For example:

{
    "monitoringCluster": "mad_dev",
    "logDirectory": "/usr/local/lib/metrics-aggregator-daemon/logs",
    "pipelinesDirectory": "/usr/local/lib/metrics-aggregator-daemon/config/pipelines",
    "httpHost": "0.0.0.0",
    "httpPort": 7090,
    "httpHealthCheckPath": "/mad/healthcheck",
    "httpStatusPath": "/mad/status",
    "jvmMetricsCollectionInterval": "PT1.0S",
    "pekkoConfiguration": {
        "pekko": {
            "loggers": [
                "org.apache.pekko.event.slf4j.Slf4jLogger"
            ],
            "loglevel": "DEBUG",
            "stdout-loglevel": "DEBUG",
            "logging-filter": "org.apache.pekko.event.slf4j.Slf4jLoggingFilter",
            "actor": {
                "debug": {
                    "unhandled": "on"
                }
            }
        }
    }
}

Pipelines

One instance of Metrics Aggregator daemon supports multiple independent services on the same host. The most basic single application host still typically configures two services: i) the end-user application running on the host, and ii) the system metrics captured by CollectD. Each of these services is configured as a pipeline in Metrics Aggregator Daemon. The pipeline defines the name of the service, one or more sources of metrics and one more destinations or sinks for the aggregated statistics.

For example:

{
    "name": "MyApplicationPipeline",
    "sources":
    [
        {
            "type": "com.arpnetworking.metrics.common.sources.ClientHttpSourceV1",
            "name": "my_application_http_source"
        },
        {
            "type": "com.arpnetworking.metrics.common.sources.FileSource",
            "name": "my_application_file_source",
            "filePath": "/var/log/my-application-query.log",
            "parser": {
                "type": "com.arpnetworking.metrics.mad.parsers.QueryLogParser"
            }
        }
    ],
    "sinks":
    [
        {
            "type": "com.arpnetworking.tsdcore.sinks.TelemetrySink",
            "name": "my_application_telemetry_sink"
        },
        {
            "type": "com.arpnetworking.tsdcore.sinks.AggregationServerHttpSink",
            "name": "my_application_aggregation_server_http_sink",
            "uri": "http://192.168.0.2:7066/metrics/v1/data"
        }
    ]
}

Each of the pipeline configuration files should be placed in the pipelinesDirectory defined as part of the daemon configuration above.

Hocon

The daemon and pipeline configuration files may be written in Hocon when specified with a _.conf extension.

Sources

Collectd

Example MAD source configuration:

{
  type="com.arpnetworking.metrics.mad.sources.MappingSource"
  name="collectd_mapping_source"
  "findAndReplace": {
    "^cpu/([\\d]+)/(cpu|percent)/([^/]+)(/value)?": ["cpu/$3", "cpu/by_core/$1/$3"],
    "^snmp/cpu_detailed/([\\d]+)/([^/]+)(/value)?": ["snmp/cpu/$2", "snmp/cpu/by_core/$1/$2"],
    "^load/load/((1|5|15)min)": ["load/$1"],
    "^memory/memory/([^/]+)(/value)?": ["memory/$1"],
    "^vmem/vmpage_number/([^/]+)(/value)?": ["vmem/$1"],
    "^vmem/vmpage_io/([^/]+)/(.*)": ["vmem/io/$1/$2"],
    "^vmem/vmpage_faults/(.*)": ["vmem/faults/$1"],
    "^swap/swap/([^/]+)(/value)?": ["swap/$1"],
    "^swap/swap_io/([^/]+)(/value)?": ["swap/io/$1"],
    "^interface/([^/]+)/if_([^/]+)/(.*)": ["interface/$1/$3/$2"],
    "^disk/([^/]+)/disk_([^/]+)/(read|write)": ["disk/$1/$3/$2"],
    "^df/(.*)(/value)?": ["disk/$1"],
    "^ntpd/(.*)(/value)?": ["ntpd/$1"],
    "^processes/ps_state/([^/]+)(/value)?": ["processes/by_state/$1"],
    "^processes/([^/]+)/ps_(vm|rss|data|code|stacksize)(/value)?": ["processes/by_name/$1/$2"],
    "^processes/([^/]+)/ps_(cputime|count|pagefaults)/(.*)": ["processes/by_name/$1/$2/$3"],
    "^processes/([^/]+)/ps_disk_([^/]+)/(.*)": ["processes/by_name/$1/disk/$3/$2"],
    "^tcpconns/([^-]+)-(local|remote)/tcp_connections/([^/]+)(/value)?": ["tcpconns/$2/$1/$3"],
    "^tcpconns/all/tcp_connections/([^/]+)(/value)?": ["tcpconns/all/$1"],
    "^memcached/df/cache/(.*)": ["memcached/cache/$1"],
    "^memcached/memcached_command/([^/]+)(/value)?": ["memcached/commands/$1"],
    "^memcached/memcached_connections/current(/value)?": ["memcached/connections"],
    "^memcached/memcached_items/current(/value)?": ["memcached/items"],
    "^memcached/memcached_octets/rx": ["memcached/network/bytes_read"],
    "^memcached/memcached_octets/tx": ["memcached/network/bytes_written"],
    "^memcached/memcached_ops/([^/]+)(/value)?": ["memcached/operations/$1"],
    "^memcached/percent/([^/]+)(/value)?": ["memcached/$1"],
    "^memcached/ps_count/.*": [],
    "^memcached/ps_cputime/.*": [],
    "^uptime/uptime(/value)?": ["uptime/value"]
  },
  "source": {
    type="com.arpnetworking.metrics.common.sources.CollectdHttpSourceV1"
    actorName="collectd-http-source"
    name="collectd_http_source"
  }
}

Example Collectd write plugin configuration:

LoadPlugin write_http
<Plugin "write_http">
  <Node "mad">
    URL "http://localhost:7090/metrics/v1/collectd"
    Format "JSON"
    Header "X-TAG-SERVICE: collectd"
    Header "X-TAG-CLUSTER: collectd_local"
    StoreRates true
    BufferSize 4096
  </Node>
</Plugin>

Telegraf

Example MAD source configuration:

{
  type="com.arpnetworking.metrics.mad.sources.MappingSource"
  name="telegraftcp_mapping_source"
  findAndReplace={
    "\\."=["/"]
  }
  source={
    type="com.arpnetworking.metrics.common.sources.TcpLineSource"
    actorName="telegraf-tcp-source"
    name="telegraftcp_source"
    host="0.0.0.0"
    port="8094"
    parser={
      type="com.arpnetworking.metrics.mad.parsers.TelegrafJsonToRecordParser"
      timestampUnit="NANOSECONDS"
    }
  }
}

Example Telegraf output configuration:

[agent]
interval="1s"
flush_interval="1s"
round_interval=true
omit_hostname=false

[global_tags]
service="telegraf"
cluster="telegraf_local"

[[outputs.socket_writer]]
address = "tcp://localhost:8094"
data_format = "json"
json_timestamp_units = "1ns"
keep_alive_period = "5m"

[[inputs.cpu]]
percpu = true
totalcpu = true
collect_cpu_time = false
report_active = false
[[inputs.mem]]
[[inputs.net]]
[[inputs.netstat]]
[[inputs.disk]]
interval="10s"
[[inputs.diskio]]
[[inputs.swap]]
[[inputs.kernel]]

Graphite

Example MAD source configuration:

{
  type="com.arpnetworking.metrics.mad.sources.MappingSource"
  name="graphitetcp_mapping_source"
  actorName="graphite-tcp-source"
  findAndReplace={
    "\\."=["/"]
  }
  source={
    type="com.arpnetworking.metrics.common.sources.TcpLineSource"
    name="graphitetcp_source"
    host="0.0.0.0"
    port="2003"
  }
}

Development

To build the service locally you must satisfy these prerequisites:

Note: Requires at least Docker for Mac Beta version Version 1.12.0-rc4-beta19 (build: 10258)

Next, fork the repository, clone and build.

Unit tests and integration tests can be run from IntelliJ. Integration tests require that the service and its dependencies be running (see below).

Building:

metrics-aggregator-daemon> ./jdk-wrapper.sh ./mvnw verify

To launch the service and its dependencies in Docker:

metrics-aggregator-daemon> ./jdk-wrapper.sh ./mvnw docker:start

To launch the service with remote debugging on port 9001 and its dependencies in Docker:

metrics-aggregator-daemon> ./jdk-wrapper.sh ./mvnw -Ddebug=true docker:start

To execute performance tests:

metrics-aggregator-daemon> ./jdk-wrapper.sh ./mvnw -PperformanceTest test

To use the local version in your project you must first install it locally:

metrics-aggregator-daemon> ./jdk-wrapper.sh ./mvnw install

You can determine the version of the local build from the pom.xml file. Using the local version is intended only for testing or development.

You may also need to add the local repository to your build in order to pick-up the local version:

  • Maven - Included by default.
  • Gradle - Add mavenLocal() to build.gradle in the repositories block.
  • SBT - Add resolvers += Resolver.mavenLocal into project/plugins.sbt.

License

Published under Apache Software License 2.0, see LICENSE

© Groupon Inc., 2014