Tags

v1.62.0

## Release notes

# 1.62.0

### ✨ New features

- **Deployment-based guardrails**  
  Enforce your own domain-specific safety logic directly using Deployments as guardrails. Whether you need to flag sensitive content, mask PII, or hard-block non-compliant requests, deployment-based guardrails give you full programmatic control and flexibility. Bring your own models, your own rules, and your own thresholds without being locked into a predefined safety framework.

- **Alerts on custom metrics**  
  Alerts can now be configured on custom metrics. Registration deployments also support custom-metric alerts. For line/bar charts an alert is raised for every x-y value pair where the y value crosses the defined threshold; for timeseries the aggregation value (sum/max/min/average) is evaluated against the threshold.

- **Data cards and model cards for repository-less deployments**  
  Data cards and model cards can now be added to deployments that do not have a linked repository.

v1.61.0

### 💥 Breaking changes
- Removed the option to change custom metric graph type

### ✨ New features
- **Time series based custom metrics**
  - A new custom metric type supports data submission over time. It supports binning similar to other metrics, with options to group submitted data within a bin using average, max, min, and sum aggregators.

v1.60.0

## Release notes

### ✨ New features

- **Reference JSON schema v2.0.0**  
  Model, explainer, and transformer references now support `schemaVersion: "2.0.0"`. The new schema introduces a more explicit structure where each Deployment backend has its own dedicated shape, replacing the flat field layout of v1. References without a `schemaVersion` field continue to be treated as v1. The default version will change to `"2.0.0"` in a future release, so migrate your existing references.

- **Bulk create custom controls from CSV**  
  Define custom controls in CSV format, convert to JSON using the Deeploy CLI (`deeploy convert-controls path/to/controls.csv`) or using the client. Then use the client to create the controls and a control framework containing those new controls.

v1.59

## Release Notes
### 🐛 Bug fixes

- Fix issue with getting the raw input and output on Request logs

v1.58.0

**Breaking changes**
- The `serverless` option has been removed in favor of `minReplicas` for Model, Explainer and Transformer objects

**Release notes**
Control how managed Deployments scale with new autoscaling options (minReplicas, maxReplicas, scaleMetric and scaleTarget). Organizations on the Private Cloud plan have access to advanced configuration options, including annotations and affinity, for even more control over deployment.

v1.57.0

Release notes

- **Added Azure Machine Learning model registry integration**: 
  - Deploy model and explainer artifacts from your Azure Machine Learning model registry on Deeploy’s deployment backend.
- Removed documentation templates from deployment flow in favour of use case

v1.56.1

Release notes

- Inference now supports streaming responses at the chat completions, completions and predict endpoints.
- Removed documentation_template_ids from Deployment creation, all documentation is now available at use cases

v1.55.0

# Release notes

✨ New features
- **Wildcard endpoints with a base endpoint** <br>
Wildcard routing now enables dynamic URL path forwarding to external Deployments, providing flexible endpoint handling for models and APIs with multiple endpoints or variable path requirements

v1.54.2

# Release notes

- Added field OidcCustomScope for deployment creation and updates

v1.54.1

## Release Notes
### 🐛 Bug fixes

- Fixed pydantic issues with creating deployments.

v1.54.0

v1.53.0

v1.52.1

v1.52.0

v1.51.0

v1.50.0

v1.49.0

v1.48.0

v1.47.1

Fix output validation for get_prediction_logs() get_request_logs() and evaluate()

v1.47.0

Release notes

✨ New features

- Custom monitoring metrics
Visualize your own metrics in the new Custom tab in Deployment monitoring. Add up to four custom metrics; calculate the data points you want to visualize and upload them with the Deeploy package