IRONdb-relay

The IRONdb-relay, like the carbon-relay or the carbon-c-relay is a metrics data router that takes carbon TEXT format metrics and routes them to the appropriate IRONdb storage node.

Since IRONdb uses SHA256 hashing to route metrics to IRONdb nodes, it is incompatible with routing options that exist in carbon-c-relay and carbon-relay. In addition, it provides advanced aggregation and filtering functions.

Features

  • Ingests TEXT carbon format metrics on a configurable port

    foo.bar.baz 1234.56 1507724786

  • Routes to primary owner of the metric name and then subsequent nodes if the primary is down.

  • Aggregation of incoming metrics based on regular expressions with support for SUM, AVG, MIN, MAX, p0, p25, p50, p95, p99, p100

  • Blacklist and whitelist filtering of metrics based on regular expressions

  • Durable delivery of metrics using write head logs

Installation

IRONdb-relay requires one of the following operating systems:

  • RHEL/CentOS, version 6.x, 7.x.
  • Ubuntu 16.04 LTS.

The following network protocols and ports are utilized. These are defaults and may be changed via configuration files.

  • 2003/tcp (Carbon plaintext submission)
  • 8112/tcp (admin UI, HTTP REST API)

System Tuning

You should follow the same system tuning as outline in the IRONdb installation.

Configure Software Sources

For EL7 or Ubuntu 16.04, use the same software source as the IRONdb installation.

For EL6, create the file /etc/yum.repos.d/Circonus.repo with the following contents:

[circonus]
name=Circonus
baseurl=http://pilot.circonus.net/centos/6/x86_64/
enabled = 1
metadata_expire = 5m

[circonus-crash-reporting]
name=Circonus - Crash Reporting
baseurl=http://updates.circonus.net/backtrace/centos/el6/
enabled = 1

Install Package

  • (EL6, EL7) /usr/bin/yum install circonus-platform-irondb-relay

  • (Ubuntu) /usr/bin/apt-get install circonus-platform-irondb-relay

Run Installer

Prepare site-specific information for setup. These values may be set via shell environment variables, or as arguments to the setup script. The environment variables are listed below.

  • IRONDB_CHECK_UUID

    (required) Check ID for Graphite metric ingestion, which must be the same on all cluster nodes. You may use the uuidgen command that comes with your OS, or generate a well-formed, non-nil UUID with an external tool or website.

  • IRONDB_CHECK_NAME

    (required) The string that will identify Graphite-compatible metrics stored in the check identified by IRONDB_CHECK_UUID. For example, if you submit a metric named "my.metric.1", and the check is named "test", the resulting metric name in IRONdb will be "graphite.test.my.metric.1".

  • IRONDB_BOOTSTRAP

    (required) The comma separated list of IRONdb nodes (ipaddress:port) to use to discover the layout of the IRONdb cluster. It's a good practice to list all IRONdb nodes in this list to adjust to down nodes.

  • IRONDB_CRASH_REPORTING

    (optional) Control enablement of automated crash reporting. Default is "on". IRONdb utilizes sophisticated crash tracing technology to help diagnose errors. Enabling crash reporting requires that the system be able to connect out to the Circonus reporting endpoint: https://circonus.sp.backtrace.io:6098 . If your site's network policy forbids this type of outbound connectivity, set the value to "off".

  • IRONDB_RELAY_DURABLE

    (optional) Control enablement of durable delivery. Default is "false". If set to "true", will cause IRONdb-relay to use the disk to persist all incoming metrics to the file system before sending them on to IRONdb nodes.

Run the setup script. All required options must be present, either as environment variables or via command-line arguments. A mix of environment variables and arguments is permitted, but environment variables take precedence over command-line arguments. Use the -h option to view a usage summary:

Usage: setup-irondb-relay [-h] -c <check-name> -u <check-uuid> -B <addr:port>[,<addr:port>,...]
       [-d] [-b (on|off)]
  -c <check-name>       : Graphite check name
  -u <check-uuid>       : Graphite check UUID
  -d                    : Use durable delivery to IRONdb
  -B <irondb-node-list> : Bootstrap to this list of IRONdb nodes
  -b on|off             : Enable/disable crash reporting (default: on)
  -h                    : Show usage summary

Example:
  setup-irondb-relay -c foo -u f2eaa1b7-f7e8-41bd-9e8d-e52d43dc88b0 -d -B 10.1.13.1:8112,10.1.13.2:8112 -b on

The setup script will configure your IRONdb-relay instance and start the service. See the Graphite Ingestion section for details.

Configuration

IRONdb-relay is implemented using libmtev, a framework for building high-performance C applications. You may wish to review the libmtev configuration documentation for an overview of how libmtev applications are configured generally.

This document deals with options that are specific to IRONdb-relay, but links to relevant libmtev documentation where appropriate.

Default values are those that are present in the default configuration produced during initial installation.

irondb-relay.conf

This is the primary configuration file that IRONdb-relay reads at start. It includes additional configuration files which are discussed later. It is located at /opt/circonus/etc/irondb-relay.conf

irondb-relay

<irondb-relay lockfile="/irondb-relay/logs/irondb-relay.lock" text_size_limit="512">

IRONdb-relay's libmtev application name. This is a required node and must not be changed.

irondb-relay lockfile

Path to a file that prevents multiple instances of the application from running concurrently. You should not need to change this.

Default: /irondb-relay/logs/irondb-relay.lock

eventer

<eventer>
  <config>
    <concurrency>8</concurrency>
    <default_queue_threads>8</default_queue_threads>
    <ssl_dhparam1024_file/>
    <ssl_dhparam2048_file/>
  </config>
</eventer>

Libmtev eventer system configuration. See the libmtev eventer documentation.

The ssl_dhparam*_file configurations are set to null to disable automatic Diffie-Hellman parameter generation at startup. IRONdb-relay does not utilize TLS by default, though this capability is present in libmtev.

logs

Libmtev logging configuration. See the libmtev logging documentation.

By default, the following log files are written and automatically rotated, with the current file having the base name and rotated files having an epoch-timestamp suffix denoting when they were created:

  • /irondb-relay/logs/errorlog: Output from the daemon process, including not just errors but also operational warnings and other information that may be useful to Circonus Support.
    • Rotated: 24 hours
    • Retained: 1 week

modules

Libmtev module configuration. See the libmtev module documentation

There are 2 modules provided with IRONdb-relay:

  • filter

    Will allow you to setup whitelist/blacklist filtering for metrics

    1. Enable the module under the <modules> section of your config by adding the line:

      <generic image="filter" name="filter_hook"></generic>

    2. Create your filter config

      Add a <filter> block to your irondb-relay.conf file. A <filter> can have exactly one <ruleset> block. A <ruleset> block can have any number of <rule> blocks. A <rule> block consists of a <match_regex> or <match_all> directive and a <result>. <rule> blocks are processed in order and processing stops at the first matching <rule>.

      Depending on whether you want a whitelist or a blacklist you would either configure your filter to whitelist a set of regexes and then have a <match_all> rule to deny everything else, or you would configure your filter to have a rule to match metrics you want to blacklist then have a final <match_all> rule to allow the remainder.

      An example of a blacklist would resemble:

       <filter> 
        <ruleset> 
          <rule> 
            <match_regex>^relay_test\.agent\.2.*</match_regex>
            <result>deny</result> 
          </rule> 
          <rule> 
            <match_all>true</match_all>
            <result>allow</result> 
          </rule> 
       </ruleset>
      </filter>
      

      The above would blacklist everything that starts relay_test.agent.2 and allow everything else.

      For best performance, it is wise to organize your <rule> blocks in descending order based on the expected frequency of matching. You want the <rule>s that match more often to be at the beginning of the list and the <rule>s that match infrequently to be lower down in the list.

  • aggregation_hook

    Will allow you to perform aggregation on incoming metrics and produce new metrics as the result.

    1. Enable the module under the <modules> section of your irondb-relay.conf by adding the line:

      <generic image="aggregation_hook" name="aggregation_hook"></generic>

    2. Create your aggregation config

      Add an <aggregation> block to your irondb-relay.conf file. An <aggregation> can have exactly one <matchers> block which itself can contain any number of <matcher> blocks. A <matcher> block consists of the following:

      • <match_regex> - the regular expression (including captures) you want to match incoming metrics
      • <flush_seconds> - how long to aggregate matching records for
      • <flush_name_template> - the template to use for the outgoing name of the aggregated metric result
      • <flush_functions> - a comma separate list of functions you want applied to the matching metric values
      • <flush_original> - whether or not you want to let the original incoming metric to also be sent to IRONdb
      • <jitter_ms> - to prevent collisions among multiple relays which might be aggregating the same metrics, set this to a unique value per irondb-relay instance
      • <idle_cycles> - how many multiples of <flush_seconds> should the relay wait before giving up on any new incoming metrics that would fall into this aggregation window

      For <flush_name_template> you can use capture references (\1) and a special sequence ${FF} to create the outgoing metric name.

      An example:

      <aggregation>
      <matchers>
       <matcher>
         <match_regex>^relay_test\.agent.[0-9]*\.metrics\.([0-9]*)</match_regex>
         <flush_seconds>10</flush_seconds>
         <flush_name_template>agg.all_agents.metrics.\1_${FF}</flush_name_template>
         <flush_functions>avg</flush_functions>
         <flush_original>false</flush_original>
         <jitter_ms>10</jitter_ms>
         <idle_cycles>2</idle_cycles>
       </matcher>
       <matcher>
         <match_regex>^relay_test\.agent.[0-9]*\.metrics\.([0-9]*)</match_regex>
         <flush_seconds>10</flush_seconds>
         <flush_name_template>foo.all_agents.metrics.\1_${FF}</flush_name_template>
         <flush_functions>sum</flush_functions>
         <flush_original>false</flush_original>
         <jitter_ms>10</jitter_ms>
         <idle_cycles>2</idle_cycles>
       </matcher>
      </matchers>
      </aggregation>
      

      The above first <matcher> matches incoming metrics that start relay_test.agent., followed by any number of digits, followed by .metrics. and finally capturing the trailing sequence of any number of digits. An example metric that would match: relay_test.agent.5.metrics.27. All incoming rows that fit this regex will have their values averaged (<flush_functions>avg</flush_functions>) for 10 seconds. The original source rows will not be sent to IRONdb and after 10 seconds a new row will be produced that looks like: agg.all_agents_metrics.27_avg and it's value will be the average of metrics.27 from all agents that this relay saw in that 10 second window.

      The 2nd <matcher> performs the same match but uses sum instead of avg and uses a different <flush_name_template>.

      The supported <flush_functions> are: sum,avg,min,max,p0,p25,p50,p95,p99,p100,histogram

      • sum is the sum of values of the matching rows.
      • avg is the mean
      • min is the smallest value
      • max is the largest
      • p0 is a synonym for min
      • p100 is a synonym for max
      • p25 is the 25th percentile value
      • p50 is the 50th percentile value
      • p95 is the 95th percentile value
      • p99 is the 99th percentile value
      • histogram is the complete distribution of all values

      With histogram IRONdb will be able to store the histogram data but there currently is no facility in graphite-web to render this data.

      A note on flushing results to IRONdb

      The very first row that matches and creates an aggregation "window" will start the flush timer. <flush_seconds> later the result will be sent to IRONdb. It is possible after this initial flush that some late data arrives that would normally fit into that same aggregation window. This is where <idle_cycles> comes into play. The relay will retain the aggregation window until no more matching rows are seen for <idle_cycles> cycles. If matching rows are seen the aggregation window is updated with the new values and the record is re-sent to IRONdb. To control the behavior of conflicts within the database when this happens please see the conflict resolver section of the IRONdb configuration.

send

This config has a single attribute: durable="true|false". If set to "true" it will use the <journal> settings below to journal every row destined for IRONdb nodes. If set to "false", it will bypass the journaling and directly send to IRONdb. If set to "false", the relay will do its best to make sure data arrives at one of the IRONdb nodes if the primary doesn't respond or is down but there is no guarantee of delivery.

listeners

Libmtev network listener configuration. See the libmtev listener documentation.

Each listener below is configured within a <listener> node. Additional listeners may be configured if desired, or the specific address and/or port may be modified to suit your environment.

Main listener

<listener address="*" port="8112" backlog="100" type="http_rest_api">
  <config>
    <document_root>/opt/circonus/share/snowth-web</document_root>
  </config>
</listener>

The main listener serves multiple functions:

Main listener address

The IP address on which to listen, or the special * to listen on any local IP address.

Default: *

Main listener port

The port number to listen on. For the main listener this will utilize both TCP and UDP.

Default: 8112

Main listener backlog

The size of the queue of pending connections. This is used as an argument to the standard listen(2) system call. If a new connection arrives when this queue is full, the client may receive an error such as ECONNREFUSED.

Default: 100

Main listener type

The type of libmtev listener this is. The main listener is configured to be only a REST API listener. This value should not be changed.

Default: http_rest_api

Graphite listener

<listener address="*" port="2003" type="graphite">
  <config>
    <check_uuid>00000000-0000-0000-0000-000000000000</check_uuid>
    <check_name>mycheckname</check_name>
    <account_id>1</account_id>
  </config>
</listener>

The Graphite listener operates a Carbon-compatible submission pathway using the plaintext format.

Multiple Graphite listeners may be configured on unique ports and associated with different check UUIDs. See the section on Graphite ingestion for details. The graphite listener config here should be kept in sync with the same configuration for the IRONdb nodes themselves.

Graphite listener address

The IP address on which to listen, or the special * to listen on any local IP address.

Default: *

Graphite listener port

The TCP port number to listen on.

Default: 2003

Graphite listener type

The type of listener. IRONdb implements a Graphite-compatible handler in libmtev, using the custom type "graphite".

Default: graphite

Graphite listener config

These configuration items control which check UUID, name, and account ID are associated with this listener. The first Graphite listener is configured during initial installation.

  • check_uuid is the identifier for all metrics ingested via this listener.
  • check_name is a meaningful name that is used in namespacing.
  • account_id is also part of namespacing, for disambiguation.

CLI listener

<listener address="127.0.0.1" port="32322" type="mtev_console">
  <config>
    <line_protocol>telnet</line_protocol>
  </config>
</listener>

The CLI listener provides a local telnet console for interacting with libmtev subsystems, including modifying configuration. As there is no authentication mechanism available for this listener, it is recommended that it only be operated on the localhost interface.

CLI listener address

The IP address on which to listen, or the special * to listen on any local IP address.

Default: 127.0.0.1

CLI listener port

The TCP port number to listen on.

Default: 32322

CLI listener type

The CLI listener uses the built-in libmtev type "mtev_console" to allow access to the telnet console.

Default: mtev_console

journal

<journal concurrency="4"
         max_bundled_messages="25000"
         pre_commit_size="131072"
/>

Journals are write-ahead logs for replicating metric data to IRONdb nodes. Each IRONdb-relay has one journal for each of the IRONdb nodes.

journal concurrency

Establishes this number of concurrent threads for writing to each peer journal, improving ingestion throughput.

Default: 4

A concurrency of 4 is enough to provide up to 700K measurements/second throughput, and is not likely to require adjustment except in the most extreme cases.

journal max_bundled_messages

Outbound journal messages will be sent in batches of up to this number, improving replication speed.

Default: 25000

journal pre_commit_size

An in-memory buffer of this number of bytes will be used to hold new journal writes, which will be flushed to the journal when full. This can improve ingestion throughput, at the risk of losing up to this amount of data if the system should fail before commit. To disable the pre-commit buffer, set this attribute to 0.

Default: 131072 (128 KB)

circonus-watchdog.conf

watchdog

<watchdog glider="/opt/circonus/bin/backwash" tracedir="/opt/circonus/traces-relay"/>

The watchdog configuration specifies a handler, known as a "glider", that is to be invoked when a child process crashes or hangs. See the libmtev watchdog documentation.

If crash handling is turned on, the glider is what invokes the tracing, producing one or more files in the tracedir. Otherwise, it just reports the error and exits.

Operations Dashboard

IRONdb-relay comes with a built-in operational dashboard accessible via port 8112 (default) in your browser, e.g., http://irondb-relay-host:8112. This interface provides real-time information about the IRONdb-relay. There are a number of tabs in the UI, which display different aspects about the node's current status.

The node's version info is displayed at top right.

Overview Tab

The "Overview" tab displays top level statistics about the relay.

Inflow

  • Socket accepts - how many connections have been made to this relay since startup
  • Received - how many individual rows have been sent into this relay
  • Parsed - the number of rows that we successfully parsed and validated
  • Parse errors - the number of parse failures
  • Aggregation gated - the number of rows that were not send on to IRONdb because of filtering or aggregation

Outflow

  • Rows sent - the number of rows sent to IRONdb nodes
  • Batches sent - rows are sent in batches, this is the count
  • Batches OK - successful batch count
  • Batch timeouts - the count of batches that timed out while sending to IRONdb nodes

Durable Delivery Tab

If <send durable="true" /> is set in the config, this tab will contain information about replication lag.

Each IRONdb node will be listed along with the number of journal reads and writes and how far behind this relay is in sending to each IRONdb node. Ideally we should have Seconds behind under 10 seconds.

Filters Tab

If you have the filter module enabled, lists each filter in your current <ruleset> and how many rows it has processed.

Aggregation Tab

If you have the aggregation_hook module enabled, lists each aggregation and how many rows it has seen, matched, skipped, and generated.

Internals Tab

Shows internal application information, such as recent error logging, job queues, open sockets, and timers. This data is used by Circonus Support when troubleshooting issues.

results matching ""

    No results matching ""