Alerts

For client integrations, Electrum creates and manages a number of configurable alerts that are broadly categorised under the following alert types:

• Tech-Ops: alerts designed to inform the relevant stakeholders of potential service issues such as service provider outages or errors relating to timeouts or declined transactions. Common alert categories include:

  • Service health alerts such as success rate alerts, failure rate alerts and timeout rate alerts.
  • Traffic alerts to flag instances where transaction processing unexpectedly drops to zero.
  • Other alerts including queue health, error rates, and others.

• Fin-Ops: alerts that advise on whether there are missing reconciliation files that prevent reconciliation and settlement from being undertaken. Common alert categories include:

  • Missing reconciliation files.

• Infrastructure and Engineering: alerts that enable Electrum to analyse the system's internal operations. Common alert categories include:

  • System resource utilisation.
  • Application performance.
  • System health.
  • Dependency health.

As a client, you can view alerts across all your integrations via the Electrum Console Alerts page:

• Currently, Electrum can provide clients with visibility of Service Health Alerts and Traffic Alerts.
• In future, Electrum intends to expand visibility to other service-related alerts, Fin-Ops alerts and Infrastructure and Engineering alerts.

Alerts Management

The Alerts Management tab lists all alerts that have been set for a client across all their integrations through Electrum in a given development environment. Alerts are grouped by service.

The tab provides an overview of the following high-level parameters pertaining to each alert:

• Enabled - indicates whether an alert is currently in operation or not by defining it as Enabled or Disabled.
• Type - indicates which operational type the alert falls under. For example, Tech-Ops, Fin-Ops, Infrastructure or Engineering. (Note that only Tech-Ops alerts are available currently.)
• Category - indicates what final transaction outcome is being alerted on. For example: timeout, success, or failure.
• Status - indicates whether an alert is currently working as intended, or not. Alert status can be designated as Active, Inactive, or Unknown.

Click on any alert in the list to reveal a drop-down menu that shows the full breakdown of the alert information, including rule details and event details.

Alerts Log

The Alerts Log tab gives a breakdown of all the alerts that have fired (i.e., triggered) within a specified time range.

This page gives a view of:

• Alerts that have triggered - the names of the alerts that have fired within the specified time range.
• Type - which operational type the alert falls under. For example, Tech-Ops, Fin-Ops, Infrastructure or Engineering. (Note that only Tech-Ops alerts are available currently.)
• Category - indicates what final transaction outcome is being alerted on. For example: timeout, success, or failure.
• Last triggered - the most recent instance of this alert firing.
• Time - the number of times the alert fired within the specified time range.

Alerts can be filtered and searched for based on the following parameters:

• Time range
• Type

Structure of Alerts

Alerts will be defined by a template structure of separate alerting parameters. These parameters are defined in (and in the future will be editable via) the Electrum Console front-end interface. The following parameters will define a full structured alert:

• Granularity - the entity/integration level at which an alert is set. Specifically, alerts can be set according to:

  • Service (e.g., Airtime, Lotto).
  • Service provider (e.g., Blue Label Telecoms, Contour).
  • Receiver (e.g., MTN, City of Cape Town municipality, Multichoice).

• Message type (specialisation) - the particular API operation or API error type.

• Transaction outcome type (category) - the outcome of a particular transaction to which the alert pertains.

  • Success
  • Decline / Failure
  • Timeout

• Threshold - the ratio, percentage, or flat count value at which an alert should fire.

• Period - the length of time time over which data will be considered for a particular alert and the frequency with which the system should check that given time period to identify whether an alert should fire.

• Equality - the mathematical definition paired with the threshold to define when an alert should fire. For example the equality and threshold can be combined to make an alert that fires when a threshold passes anything greater than 90%.