Add resource to manage ML anomaly detection jobs #1329
Conversation
| ), | ||
| }, | ||
| { | ||
| Config: testAccResourceAnomalyDetectionJobComprehensive(jobID), |
There was a problem hiding this comment.
Would it be worth separating into a basic create / update and "comprehensive" create / update test so we get coverage for things like updating the minimal example and creating the comprehensive example?
| } | ||
|
|
||
| // Convert API response back to TF model | ||
| diags.Append(job.fromAPIModel(ctx, &response.Jobs[0])...) |
There was a problem hiding this comment.
Would it be worth adding an error diag here if there is more than one job in response.Jobs? This would probably only be future proofing against the read API call being updated without the corresponding change here.
| stringplanmodifier.RequiresReplace(), | ||
| }, | ||
| Validators: []validator.String{ | ||
| stringvalidator.LengthBetween(1, 64), |
There was a problem hiding this comment.
Is this really capped at 64 characters? https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-ml-put-job#operation-ml-put-job-job_id
I didn't see this in the docs
| Attributes: map[string]schema.Attribute{ | ||
| "bucket_span": schema.StringAttribute{ | ||
| MarkdownDescription: "The size of the interval that the analysis is aggregated into, typically between 15m and 1h. If the anomaly detector is expecting to see data at near real-time frequency, then the bucket_span should be set to a value around 10 times the time between ingested documents. For example, if data comes every second, bucket_span should be 10s; if data comes every 5 minutes, bucket_span should be 50m. For sparse or batch data, use larger bucket_span values.", | ||
| Required: true, |
There was a problem hiding this comment.
Is this actually required? In the request body docs it makes it seem not required: https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-ml-put-job#operation-ml-put-job-body-application-json
| }, | ||
| NestedObject: schema.NestedAttributeObject{ | ||
| Attributes: map[string]schema.Attribute{ | ||
| "function": schema.StringAttribute{ |
There was a problem hiding this comment.
Is this actually required?
| Optional: true, | ||
| Computed: true, | ||
| Validators: []validator.String{ | ||
| // TODO: TB If only json is supported then why are we allowing delimited? |
2 participants
Fixes #520
This resource only manages the existence of an anomaly detection job, it does not manage the state (open/closed). I'm planning on adding another resource to purely manage that state, it's an independent API. IMO it's simpler for practitioners if these are separate resources altogether.