Comment on page

New Relic

New Relic One is an observability platform, offering support from monoliths to serverless. It helps engineers to instrument, analyze, troubleshoot and optimize the entire software stack - all from one place.
To integrate steadybit with New Relic, there are two general possibilities:
New Relic to steadybit: checking for specific events from New Relic in steadybit during experiment execution
steadybit to New Relic: pushing steadybit events during experiment execution to New Relic
New Relic to steadybit
We are able to react to the different events of New Relic and process them within the execution of an experiment. You can filter the events and determine at which event type the experiment should fail.
Configuration
Under Settings/Monitoring Integrations/New Relic you can enable the State Check by providing the New Relic Rest API Base URL and the REST API token that was previously created in New Relic. That's all we need to react on events of New Relic.
Usage
Once New Relic has been configured in the platform's settings, you can use the New Relic State Check in an experiment. Detailed explanation can be found in Integrate / Monitoring / New Relic.
steadybit to New Relic
By connecting steadybit to New Relic Insights you are able to investigate the effect of an attack within your familiar monitoring setup (e.g. New Relic dashboards). Thus you can see the direct effect of an attack to certain metrics like erroneous calls, CPU usage or whatever metric is important to you.
Configuration
Within steadybit go into Settings/Monitoring Integrations/New Relic, where you can enable the sending of events by providing the Insights Base URL, New Relic Account Id as well as the Insights Insert Key. The former one can be looked up in the New Relic documentation, the latter two are created in the New Relic settings upfront.
Usage
As soon as the configuration of New Relic Insights is provided, steadybit sends events of every executed experiment automatically to New Relic. These events can be configured for instance as an own widget on your dashboard by using NRQL (New Relic Query Language). An example is given below as well as a short explanation of the data structure.
Example Widget
Below you see an example widget that can be configured on your New Relic dashboard. This helps you to correlate the execution of an experiment with other dashboard widgets.
New Relic Insights Example
Events of steadybit are published as the custom event ExperimentExecution which can be queried using NRQL. Therefore, go within New Relic to the Dashboard, add a new widget of type chart and use for instance the following NRQL:
SELECT count(targetName)
FROM ExperimentExecution
TIMESERIES FACET
    cases(  WHERE attack IN ('container-cpu-attack', 'host-cpu-attack') AS 'cpu',
            WHERE attack IN ('container-memory-attack', 'host-memory-attack') as 'memory',
            WHERE attack in('container-io-attack', 'host-io-attack') as 'IO'),
    container.host
Data Structure
Experiment Events coming from steadybit have generally the structure below. Thereby, executionType covers what was exectued (e.g. attack or action for a load test) and status whether it was started or stopped.
Attribute
Description
Values
experimentKey
Key of the experiment (consisting of teamname and number)
e.g. BS-123
executionId
Unique number of this specific execution instance
e.g. 1020
executionStep
Indicates the executed step of the experiment or empty, if it is the start or end of the experiment
EMPTY, action or attack
executionStatus
Indicates the executed step of the experiment or empty, if it is the start or end of the experiment
EMPTY, action or attack
In addition, events covering the execution of an Attack (executionTyp = attack) have additional attributes:
Attribute
Description
Values
attack
Attack being performed
e.g. container-cpu-attack, host-cpu-attack, container-stop-attack
targetType
Type of the target
com.steadybit.extension_host.host, com.steadybit.extension_container.container, com.steadybit.extension_jvm.application
targetName
Name of the target
e.g. k8s_fashion-bestseller_fashion-bestseller-bb896d756..., ip-10-2-3-122
k8s.container.name)
Name of the container
e.g. fashion-bestseller
container.host
Name of the target
e.g. ip-10-2-3-122
...
All remaining attributes available also in the discovery and design of experiments​
...
Useful Resources
​EU and US Region data centers​
​API Keys​

Attribute	Description	Values
`experimentKey`	Key of the experiment (consisting of teamname and number)	e.g. `BS-123`
`executionId`	Unique number of this specific execution instance	e.g. `1020`
`executionStep`	Indicates the executed step of the experiment or empty, if it is the start or end of the experiment	EMPTY, `action` or `attack`
`executionStatus`	Indicates the executed step of the experiment or empty, if it is the start or end of the experiment	EMPTY, `action` or `attack`

Attribute	Description	Values
`attack`	Attack being performed	e.g. `container-cpu-attack`, `host-cpu-attack`, `container-stop-attack`
`targetType`	Type of the target	`com.steadybit.extension_host.host`, `com.steadybit.extension_container.container`, `com.steadybit.extension_jvm.application`
`targetName`	Name of the target	e.g. `k8s_fashion-bestseller_fashion-bestseller-bb896d756...`, `ip-10-2-3-122`
`k8s.container.name`)	Name of the container	e.g. `fashion-bestseller`
`container.host`	Name of the target	e.g. `ip-10-2-3-122`
...	All remaining attributes available also in the discovery and design of experiments	...

Instana

Prometheus

Last modified 3mo ago