Deliver Deeper Operational Intelligence: Introducing Alert Functions Library Through SignalFx API

Monitoring cloud applications requires ingesting and analyzing data from hundreds to thousands of web services, many of which employ scale-out, elastic architectures built on cloud infrastructure, containers, and applications. Traditional approaches to setting alerting conditions have been confined to using charts to make sure things look as expected and creating an alert to notify someone when action is required. However, today’s cloud environments are dynamically changing and cloud operations teams require more flexibility in alerting for their operational use cases.

SignalFx’s new Alert Functions Library enables cloud operations to develop custom alerts that reflect the needs of their environment. With the SignalFx API, express the most advanced alert logic for a wide range of use cases. Deliver deeper operational intelligence in a much easier way with direct access to the SignalFx alert conditions library.

Programmatically Address Operational Issues

The SignalFx API provides access to Built-in Alert Conditions and additional alert functions for cloud operations teams to incorporate alerts anywhere in the operational workflow. Full access to alert functions means that development and operations teams can leverage pre-packaged alerting algorithms to use in any custom alert they need to develop. Just as Built-in Alert Conditions gives you a starting point accessible through the SignalFx UI, the Alert Conditions Library gives the same starting point and even more flexibility to generate meaningful alerts threshold. Read more about new Built-in Alert Conditions.

An alert using built-in alert conditions is created and updated the same way other detectors are with the /v2/detector API. Define the alert condition required and the parameters to pass to the conditions based on your specific use case. Once decided, simply POST the alert with the SignalFlow program text created. The API below allows stamping out new alert detectors that leverage built-in alert conditions, for example, one per team.

# Consider using detector/validate before posting your detector
# See https://developers.signalfx.com/v2/docs/detectorvalidate
$ curl \
  --request POST \
  --header "X-SF-TOKEN: YOUR_ACCESS_TOKEN" \
  --header "Content-Type: application/json" \
  --data-binary @- \
  https://api.signalfx.com/v2/detector << EOF
{
  "name": "CPU load too high",
  "programText": "from signalfx.detectors.population_comparison\ncpu = data(‘cpu.utilization’)\npopulation.detector(cpu).publish('CPU different')",
  "rules": [{
    "severity": "Critical",
    "detectLabel": "CPU different",
    "notifications": [{ "type": "Email", "email": "person@example.com" }]
  }]
}
EOF

Alerting for New Operational Use Cases

The availability of the SignalFx API opens up access to both the data collected and the power of the SignalFx analytics engine. Rather than limiting the use to either a chart in a UI or a mechanism to alert someone, new operational use cases can include cluster management in a microservices system to automatic remediation of one-off operational anomalies and rolling restarts in a distributed environment.

Alerting in these new operational use cases requires the flexibility to describe any complex computation. Customize when alert conditions are triggered and how action is taken, whether manually or automatically with incident management tools. API access to the underlying data of any cloud environment and to a rich library of alert functions mean cloud operations can take this functionality anywhere.

The Alert Functions Library ultimately gives you a starting point to generate meaningful alert thresholds and additional flexibility to embed alert thresholds across your operational workflow. Programmatic access to pre-built alerting models enables deeper operational intelligence for effectively monitoring and operation their cloud applications.

Join us for a webinar on Optimal Alerting in Cloud Environments featuring SignalFx customer Acquia »

Thanks,
Jessica Feng