Service-level management¶

Service-level management (SLM) defines and monitors the quality of the IT services provided. It is an ITIL process discipline whose goal is to bring the performance of IT services permanently in line with customer expectations.

The implementation is based on the approach of Google: SRE fundamentals 2021: SLIs vs SLAs vs SLOs

Service-level management in Versio.io offers the following functionalities:

• Definition of service-level agreements (SLA) to group service level objectives.
• Definition and identification of service-level objectives (SLO) to determine service quality in terms of availability and performance (latency).
• Visualization of SLO status over time as well as their metrics (error budget) in a dashboard.
• Accounting of penalty payments based on SLO definitions and SLAs.

Service-level agreements and service-level objectives are stored in Versio.io in the Asset & Configuration Item Inventory. This makes standard functionalities of the Versio.io platform, such as reporting or IT governance, available and allows the data to be used by customers via the OneAPI.

Service-level agreements¶

Service-level agreements represent a grouping element for service-level objectives. The grouping can be used as a filter in the dashboard and as a container for the penalty calculation.

SLOs are assigned to one or more SLAs in the SLO importer configuration.

Figure: Service-level agreement editor

Service-level objectives¶

Configuration¶

Service-level objectives can be defined in Versio.io as an inventory configuration. Here, availability and performance can be defined as quality criteria. Based on the configuration, the SLOs are then continuously recorded by OneGate. All details about the SLO configuration can be found here.

The following availability and latency (performance) measurements are supported:

• Supported
  • HTTP requests for surveying web-based applications and APIs.
• Support planned (customer requests welcome)
  • HTTP requests for surveying content of web-based APIs
  • PING for surveying IP-based devices and components (host, virtual machine, network device, etc.)
  • SSH incl. login for measuring SSH services (host, virtual machine, network device, etc.)
  • Database queries for the measurement of database management systems
  • Database queries for surveying database contents
  • SSH for measuring metrics of operating systems (host, virtual machine, network device, etc.)

Dashboard¶

The service-level management dashboard provides users with an overview of the status over time and the control-relevant metrics of the individual SLOs as well as their totality, depending on a freely selectable time period.

Strong filtering capabilities allow the dashboard to be filtered to a specific SLA or SLO, or an individual time period under consideration. Please note that all metrics (availability, error budget) are always calculated on the basis of the observation period!

Figure: Service-level management dashboard

The following table shows which status an SLO can assume in its lifetime and how this is evaluated by the assessment engine:

Table: SLO status and interpretation

The error budget is calculated based on the SLO target and the period under analysis. The following representations are supported: * Error budget % = 100% - SLO target * Error budget time = 100% - SLO target * Observation period * Overdrawn error budget time = 'NOT GOOD' rating time - Error budget time

The analysis period must be taken into account in the dashboard, since e.g. status and costs are always calculated for the specific analysis period!

Penalty billing¶

All SLOs with the status 'NOT RELEVANT' are not considered in the evaluation. This means that the period considered is reduced accordingly. The error costs are only included for SLOs that have exceeded the allowed error costs and are calculated as follows:

Error costs = Overdrawn error budget time * Cost per time unit