1.2.x

Introduces deeper support for release automation with a wider range of HTTP traffic metrics and the integration of release plans into the Release Agents.

HTTP Metrics

Additional request count and response timemetrics have been added. These are grouped by HTTP request type to allow policy conditions to be set for GET, POST, PUT and DELETE etc requests.

This provides improved support for the SRE concept of Service Level Objectives (SLOs), specifically the monitoring of latency-based SLOs.

Release Plans

The release plans have been moved into Vault and release plan based triggers have been added to the Release Agents.

In previous versions, release plans were external artifacts stored in git repos and using them required adding additional scripts to your CD pipeline.

Key Changes

• Changed the default Pulse event type from elasticsearch to nats

• Changed Lifter to deploy NATS Streaming Server as a dependency

• Added request count and response timemetrics grouped by HTTP request type

• Added a NATS streaming channel for release plan events

• Fixed /events API endpoint returns 503 error when Elasticsearch index doesn't exist

• Fixed allocation metrics show minAvailable: true when a Deployment has no Pods (Kubernetes)

• Fixed Lifter does not create Service Account for the Quantification workflow (Kubernetes)

1.1.x

The internal, shared event queue has been replaced with individual, external event queues based on the NATS Streaming.

The internal event queue did not support durable subscriptions and the event delivery to the Workflows could fail, most noticeably immediately after a restart.

Key Changes

• Changed Kubernetes livenessProbe to include all API endpoints

• Changed the health metric on Kubernetes to include a delta of the Pod restart count

• Added NATS Streaming Server as a required component

• Added NATS Streaming as a Pulse event type

• Fixed multiple Pulse events are triggered when editing a Gateway in UI

• Fixed VGA templates cannot be updated through the Lifter UI

1.0.x

Optimised the MySQL connection management to improve startup time.

​