how to setup output bucket ?

[Alexandre Grais]

So far I am having my model and a training which was ok.

Now when forecasting i got the following issue:

loudml_1             | 172.20.0.35 - - [2020-05-27 15:57:59] "POST /models/org_cpu_mean_usage_idle__1m/_forecast?from=2020-05-27T15:57:00.000Z&to=2020-05-30T15:57:00.000Z&save_output_data=true&output_bucket=loudml&bg=true HTTP/1.1" 202 153 0.074071
loudml_1             | 172.20.0.35 - - [2020-05-27 15:58:04] "GET /jobs/df5c9d55-9d97-4307-ba3d-01b89c36943d HTTP/1.1" 200 359 0.000948
loudml_1             | 172.20.0.35 - - [2020-05-27 15:58:09] "GET /jobs/df5c9d55-9d97-4307-ba3d-01b89c36943d HTTP/1.1" 200 359 0.000628
loudml_1             | 172.20.0.35 - - [2020-05-27 15:58:09] "GET /models/kapptivate_kcollector_cpu_mean_usage_idle__1m HTTP/1.1" 200 770 0.002862
loudml_1             | INFO:schedule:Running job Every 1 minute do daemon_clear_jobs() (last run: 2020-05-27 15:57:11, next run: 2020-05-27 15:58:11)
loudml_1             | INFO:root:job[df5c9d55-9d97-4307-ba3d-01b89c36943d] forecasted values for 865 time buckets
loudml_1             | ERROR:root:cannot load bucket: Bucket not found (name = 'loudml')

I was wondering where the output bucket should be configured to change the name to the existing bucket I already have.

[Alexandre Grais]

for information I am using the modified chronograf to have the create baseline button. I think I saw in the API specs that we can specify the output_bucket in the API call.

[Sebastien Leger]

Hello, bonjour Alexandre,

Use the latest docker-compose example available in develop branch, directory docker/compose. Static configuration file defines the loudml bucket here:

loudml.yml

Use loudml/loudml:nightly image tag for now (until I release 1.6.0 tag).

If you are using the command line, I published a REPL command line tool that can be installed via PyPi
The tool has a command line flag to save data points to a given bucket.