Service Level Objectives and Agreements
Service Level Indicators (SLIs) are the metrics used to measure the level of service provided to end users (e.g., availability, latency, throughput). Service Level Objectives (SLOs) are the targeted levels of service, measured by SLIs. Monitoring and alerting will be performed by using Prometheus and Grafana
| Indicator | Objective | |
|---|---|---|
| Aggregated Availability | the percentage of requests served successfully, i.e. 100 * (total requests served successfully 2XX status code / total valid requests 2XX/5XX status code) through the internal API Gateway | >99% of requests to POST /events result in a status 200 (not 500 error) |
| Endpoint response time | Response time for all API requests, including success and failure | 99th centile of duration in a 24 hour window < 1,000ms |
| End to end latency | Time from event published by a supplier to being consumed by an acquirer | 99th centile of duration in a 24 hour window < 1h |
| Service latency | Time between an event being ingested to being available for an acquirer | 99th centile of duration in a 24 hour window < 5m |
| Ingestion availability | This will be defined per data source. For example with HMPO, is the file available within the agreed window? | - |
| Target Capacity | The service is able to meet the above SLAs with load not exceeding the target number of events in a given time window | 6,000 events in a 24 hour window |
| Data Subject Requests | All requests for supporting Access Requests from relevant Data Controllers will be handled within a given time frame | 7 days |
Service Capacity
As above, the service is designed for a capacity approximately twice the expected peak load, and tested to such. In non production environments, the mock data generation averages 5 events per minute, or 7,200 in 24 hours, but gradually released through the day, so is not representative of a peak load. If required additional peak load testing data can be arranged.
Service Availability (RPO, RTO)
In addition to the Aggregated Availability and other SLOs above, we also have targets around RPO/RTO.
During the private beta, we’re aiming for
- restoration attempted during business hours, which may be 72 hours (over a weekend). Further details in the incident levels in incident managment.
- loss of up to one hour of metadata/audit data, but acquirers will receive all data from upstream suppliers, potentially with delay