Plugins

All plugins are configured within the plugins section of the TARDIS configuration. Using multiple plugins is supported by using a separate MappingNode per plugin.

Plugins:
    Plugin_1:
        option_1: my_option_1
    Plugin_2:
        option_123: my_option_123

SQLite Registry

The SqliteRegistry implements a persistent storage of all Drone states in a SQLite database. The usage of this module is recommended in order to recover the last state of TARDIS in case the service has to be restarted.

Available configuration options

Option	Short Description	Requirement
db_file	Location of the SQLite database.	Required

Example configuration

Plugins:
  SqliteRegistry:
    db_file: drone_registry.db

Telegraf Monitoring

The TelegrafMonitoring implements an interface to monitor state changes of the Drones in a telegraf service running a UDP input module.

Note

By default the machine name of the host running tardis is added as default tag. It can be overwritten by adding tardis_machine_name: ‘something_else’ as default_tag in the configuration.

Available configuration options

Option	Short Description	Requirement
host	Hostname or IP address the telegraf UDP input module is listening to.	Required
port	Port the telegraf UDP input module is listening on.	Required
default_tags	Tags that should be included by default for all entries sent to telegraf.	Optional

Example configuration

Plugins:
  TelegrafMonitoring:
    host: der_telegraf.foo.bar
    port: 8092
    default_tags:
      something_default: 'The Default Tag'

Prometheus Monitoring

The PrometheusMonitoring implements an interface to monitor the number of drones in the states Booting, Running, Stopped, Deleted, and Error.

Available configuration options

Option	Short Description	Requirement
addr	Address on which the metrics are served on for Prometheus	Required
port	Port on which the metrics are served on for Prometheus	Required

Example configuration

Plugins:
  PrometheusMonitoring:
    addr: 127.0.0.1
    port: 8080

ElasticSearch Monitoring

The ElasticsearchMonitoring implements an interface to push the done state to an Elasticsearch instance at every state change.

Available configuration options

Option	Short Description	Requirement
host	Hostname or IP address of the Elasticsearch instance.	Required
port	Port the Elasticsearch instance is listening on.	Required
index	Target index in the Elasticsearch instance	Required
meta	Additional meta data (can be used to distinguish TARDIS instances).	Optional

Example configuration

Plugins:
  ElasticsearchMonitoring:
    host: elasticsearch.foo.bar
    port: 9200
    index: cobald_tardis
    meta: instance1

Auditor Accounting

The Auditor implements an interface to push information from the drones relevant for accounting to an Auditor instance. The plugin extracts the components to be accounted for from the MachineMetaData in the configuration. Scores which help relating resources of the same kind with different performance to each other can be configured as well. Scores are configured for each MachineType individually and multiple scores per MachineType are possible. An Auditor record requires a site_id, a user_id and a group_id. The latter two can be configured in the Auditor plugin configuration (and default to tardis if omitted). The site_id is taken from the Sites in the TARDIS config.

Available configuration options

Option \| Short Description			Requirement
host \| Hostname or IP address of the Auditor instance.			Required
port \| Port on which the Auditor instance is listening on.			Required
user \| User name added to the record. Defaults to `tardis`.			Optional
group \| Group name added to the record. Defaults to `tardis`.			Optional
components \| Configuration of the components per `MachineType`. Used to attach scores to individual components.			Optional

Example configuration

Plugins:
  Auditor:
    host: "127.0.0.1"
    port: 8000
    user: "some-user"
    group: "some-group"
    components:
      machinetype_1:
        Cores:
          HEPSPEC06: 1.2
          OTHERBENCHMARK: 1.4
      machinetype_2:
        Cores:
          HEPSPEC06: 1.0
        Memory:
          PRECIOUSMEMORY: 2.0

Your favorite monitoring is currently not supported? Please, have a look at how to contribute.