Toni Moreno

Donut Model doesn't fit well CPU usage , score always 100%

3

Hello. I've readed loudml documentation and also some examples I've found , and I cannot understand, how I should create/train the model to fit well my example metric "usage_system cpu"...

Hi @regel , I've configured some scheduled jobs from my loudml-config.yml ``` scheduled_jobs: - name: "forecast({{model_name}}) every five minutes" relative_url: "/models/{{model_name}}/_forecast" method: post params: from: "now-1h" to: "now+1h" save_output_data: True...

[Error on create model] "invalid field features.0.measurement: not a valid value" (name limited by config "^[a-zA-Z0-9-_@]+$") ??

3

Hello @regel , sorry for this new issue. I'm trying to create a model with this definition. ``` { "bucket_interval":"1m", "default_bucket":"output-influx1", "features":[ { "anomaly_type":"low_high", "default":0, "field":"value", "io":"io", "match_all":[ { "tag":"metric_id",...

[Error] - socket.timeout: timed out | During handling of the above exception, another exception occurred | TypeError: __init__() takes from 1 to 2 positional arguments but 4 were given

2

Hi @regel , I have one loudml server working with 7 models ( testing which one better fit the metric shape). Sometimes , in the log the following message is...

[Trainning Failed] hyperopt.exceptions.AllTrialsFailed

1

Hello again , and sorry for the inconvenience When trying to re-train one of my models, sometimes they are failing with this message, and I can not understand what exact...

[Feature Request] Prepend Timestamp and JobID to the output logs.

1

Hello @regel . Could be great if output log lines had timestamp to see when the output happened. Y have deployed loudml in a swarm and has been restarted several...

Just after begin trainning for 7 different models. ``` docker@worker3:~$ docker exec -it 5294608ad09b loudml Connected to localhost:8077 version 1.6.0 Loud ML shell 1.6.3 > list-jobs Traceback (most recent call...

Related to this comment https://github.com/regel/loudml/issues/375#issuecomment-669721615 ``` > list-model-versions swarm@cpu@mean@usage_active@host_worker2_cpu_cpu-total@time@1m version active loss trained 00 0. 116.357 1.

Related to this 2 issues: https://github.com/vsergeyev/loudml-grafana-app/issues/21 https://github.com/regel/loudml/issues/62 In the context of automation/visualization would be great is we could parse certain parameters from model name. Right now as indicated here https://github.com/regel/loudml/issues/62...

[Feature Request] Add ability to Autoasign port in server and API to check available endpoints

1

Use case: This feature request could be useful on cloud environments, where docker images could be created and ended. In this context. Supose you have two docker images starting a...