Elasticsearch Integration

Horizon can persist and forward certain data to Elasticsearch. The following sections describe the configuration possibilities as well as the available features. Internally, all Elasticsearch integrations use the Jest library to access the Elasticsearch REST interface.

Configuration

The location of the configuration file depends on the feature.

For flows, find it in $OPENNMS_HOME/etc/org.opennms.features.flows.persistence.elastic.cfg. You can set the following properties:

Property Description Default

Required

elasticUrl

URL(s) to Elasticsearch nodes. Can either point directly to REST API or seed nodes. The format is: <host>:<port>. Comma separate multiple values.

http://localhost:9200

Optional

elasticIndexStrategy

Index strategy for data. Allowed values: yearly, monthly, daily, hourly.

daily

globalElasticUser

Username to use for all nodes, when X-Pack Security is configured.

none

globalElasticPassword

Password to use for all nodes, when X-Pack Security is configured.

none

ignoreCertificates

Set this to ignore HTTPS/SSL/TLS certificates.

false

defaultMaxTotalConnectionPerRoute

Sets the default maximum connections per route. Negative values are ignored.

<available processors> * 2

maxTotalConnection

Sets the default maximum total connections. Negative values are ignored.

<max connections per route> * 3

nodeDiscovery

Enable/disable node discovery. Valid values are true|false.

false

nodeDiscoveryFrequency

Defines the frequency, in seconds, in which the nodes are rediscovered. Must be set, if discovery=true

none

proxy

Defines a proxy server for communicating with your Elasticsearch cluster. Must be a valid URL.

none

httpCompression

Defines that HTTP compression is enabled.

none

retries

Defines how many times an operation is retried before considered failed.

0

retryCooldown

Defines the cool-down in ms to wait before retrying. Value of 0 means no cool-down. Value must be >= 0.

500

connTimeout

Defines the connection timeout in ms.

5000

readTimeout

Defines the read timeout in ms.

30000

bulkRetryCount

Defines the number of retries performed before a bulk operation is considered failed. When bulk operations fail, only the failed items are retried.

5

bulkSize

The number of flow documents to collect into a bulk operation before committing. This is per thread. Set to 0 to disable bulking.

1000

bulkFlushMs

Timeout to flush bulk even if bulkSize wasn’t reached. This is per thread. Set to 0 to disable flushing.

500

settings.index.number_of_shards

The number of primary shards that an index should have. Refer to Elasticsearch Reference → Index Modules for more details.

none

settings.index.number_of_replicas

The number of replicas each primary shard has. Refer to Elasticsearch Reference → Index Modules for more details.

none

settings.index.refresh_interval

How often to perform a refresh operation, which makes recent changes to the index visible to search. Refer to Elasticsearch Reference → Index Modules for more details.

none

settings.index.routing_partition_size

The number of shards a custom routing value can go to. Refer to Elasticsearch Reference → Index Modules for more details.

none

indexPrefix

Prefix is prepended to the index and template names. Used in cases where you want to share the same Elasticsearch cluster with many Horizon instances.

none

circuitbreaker.failure_rate_threshold_in_percent

The failure rate threshold in percentage above which the CircuitBreaker should trip open and start short-circuiting calls.

99

circuitbreaker.wait_in_open_state_milliseconds

The time, in milliseconds, that the CircuitBreaker should wait before transitioning from open to half-open.

60000

circuitbreaker.ring_buffer_size_in_closed.state

The size of the ring buffer when the CircuitBreaker is closed. The ring buffer needs to be filled before the failure rate is calculated.

100

circuitbreaker.ring_buffer_size_in_half_open_state

The size of the ring buffer when the CircuitBreaker is half-open. This ring buffer is used when the breaker transitions from open to half-open to decide whether the circuit is healthy or not.

10

If a configuration management tool is used, the properties file can be created and used as the startup configuration.
If credentials are provided, preemptive authorization is used for all defined Elasticsearch nodes.
Configuration example to access Elasticsearch
elasticUrl=http://elastic:9200
elasticIndexStrategy=daily
globalElasticUser=elastic
globalElasticPassword=changeme

Credentials

It is possible to define credentials for each Elasticsearch node individually. Credentials for each node must be stored in $OPENNMS_HOME/etc/elastic-credentials.xml.

Custom credentials
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<elastic-credentials>
    <credentials url="http://localhost:9200" username="ulf" password="ulf" />
    <credentials url="https://10.10.0.1:9333" username="ulf" password="flu" />
</elastic-credentials>
Credentials are globally defined and will be used by all features that connect to the specified URLs.

Circuit breaker

When Elasticsearch is unavailable, a circuit breaker is used to reduce the number of blocked threads. After a defined number of errors, requests to the server will not be sent for the defined period of time. Instead, the requests are cancelled immediately with an error. To configure the circuit breaker, use the parameters that begin with circuitbreaker..

Further documentation can be found at here