Updating endpoint with a configuration file
Verta's client and CLI support updating Endpoint with JSON or YAML configuration file. This tutorial will show how it can be done, using the client and CLI.
Sometimes, it is more convenient to specify update configuration in a separate, human-readable file, than to have everything in a Python script or a split CLI command.
For example:
{
"run_id": "<experiment run id>",
"strategy": "canary",
"canary_strategy": {
"progress_step": 0.05,
"progress_interval_seconds": 30,
"rules": [
{"rule": "latency_avg_max", "rule_parameters": [{"name": "threshold", "value": 0.1}]},
{"rule": "error_4xx_rate", "rule_parameters": [{"name": "threshold", "value": 1}]}
]
},
"autoscaling": {
"quantities": {"min_replicas": 0, "max_replicas": 4, "min_scale": 0.5, "max_scale": 2.0},
"metrics": [
{"metric": "cpu_utilization", "parameters": [{"name": "target", "value": 0.5}]},
{"metric": "memory_utilization", "parameters": [{"name": "target", "value": 0.7}]}
]
},
"env_vars": {"VERTA_HOST": "<Verta host URL>"},
"resources": {"cpu": 0.25, "memory": "100M"}
}
The JSON representations for endpoint configuration values can be found in their respective Python API documentation entries.
This allows for easier inspection and modification later on. Users can find out more about the fields of the configuration file here.
With the client, instead of using
Endpoint.update(), we will use the Endpoint.update_from_config() method, which takes only the path to the configuration file:endpoint.update_from_config("config_file.json")
Last modified 8mo ago