Anomaly Detection Plugin stuck in Init

I am using the anomaly detection plugin (v7.8.0). I am able to create a detector and can preview anomalies from within the Feature creation section.
However, when I start the detector, it remains stuck in the Initializing state.

In the .opendistro-anomaly-results-history index I keep seeing this error:

Threshold model node is not available.

Can someone help with this?

Thank you.

Elasticsearch versions: 7.8.0
Kibana version: 7.8.0

Could you run

GET /_cat/nodes

and share the output here?

Yeah, here it is:

ip             heap.percent ram.percent cpu load_1m load_5m load_15m node.role master name
100.96.3.136             10          95  39    0.40    0.36     0.41 lr        -      elasticsearch-logstash-clients-3
100.96.12.162            62          93 100   10.86   12.51    12.26 dilrt     -      elasticsearch-data-2a-1
100.96.1.16              49          96   0    0.17    0.09     0.03 mr        -      elasticsearch-master-1
100.96.11.193            65          84 100   11.26   12.98    12.54 dilrt     -      elasticsearch-data-2a-5
100.118.37.177           24          96   4    0.21    0.13     0.11 mr        *      elasticsearch-master-2
100.96.12.233            27          99 100   12.15   10.28    10.76 dilrt     -      elasticsearch-data-2a-3
100.96.20.3              19          92 100    9.82   10.95    10.79 dilrt     -      elasticsearch-data-2a-0
100.96.1.32              60          95  41    0.38    0.68     0.69 lr        -      elasticsearch-logstash-clients-2
100.96.1.132             27          96   2    0.10    0.10     0.13 lr        -      elasticsearch-client-0
100.96.0.5               47          96 100   10.93   11.08    11.02 dilrt     -      elasticsearch-data-2a-4
100.96.11.192            57          95  50    1.47    1.24     1.03 lr        -      elasticsearch-logstash-clients-0
100.96.38.96             53          96 100   15.39   13.46    11.83 dilrt     -      elasticsearch-data-2a-2
100.96.0.16              36          95  56    0.24    0.47     0.59 lr        -      elasticsearch-logstash-clients-1
100.96.3.144             48          96   0    0.08    0.03     0.02 mr        -      elasticsearch-master-0

Can you also run the following command?

GET /_cat/nodeattrs?v

Does your node attributes contain “box_type”? Are you using hot or warm nodes?

Sure, here is the output:

node                             host           ip             attr              value
elasticsearch-master-2           100.118.37.177 100.118.37.177 xpack.installed   true
elasticsearch-master-2           100.118.37.177 100.118.37.177 transform.node    false
elasticsearch-master-1           100.96.1.16    100.96.1.16    xpack.installed   true
elasticsearch-master-1           100.96.1.16    100.96.1.16    transform.node    false
elasticsearch-master-0           100.96.3.144   100.96.3.144   xpack.installed   true
elasticsearch-master-0           100.96.3.144   100.96.3.144   transform.node    false
elasticsearch-logstash-clients-3 100.96.3.136   100.96.3.136   ml.machine_memory 13958643712
elasticsearch-logstash-clients-3 100.96.3.136   100.96.3.136   ml.max_open_jobs  20
elasticsearch-logstash-clients-3 100.96.3.136   100.96.3.136   xpack.installed   true
elasticsearch-logstash-clients-3 100.96.3.136   100.96.3.136   transform.node    false
elasticsearch-logstash-clients-2 100.96.1.32    100.96.1.32    ml.machine_memory 13958643712
elasticsearch-logstash-clients-2 100.96.1.32    100.96.1.32    ml.max_open_jobs  20
elasticsearch-logstash-clients-2 100.96.1.32    100.96.1.32    xpack.installed   true
elasticsearch-logstash-clients-2 100.96.1.32    100.96.1.32    transform.node    false
elasticsearch-logstash-clients-1 100.96.0.16    100.96.0.16    ml.machine_memory 13958643712
elasticsearch-logstash-clients-1 100.96.0.16    100.96.0.16    ml.max_open_jobs  20
elasticsearch-logstash-clients-1 100.96.0.16    100.96.0.16    xpack.installed   true
elasticsearch-logstash-clients-1 100.96.0.16    100.96.0.16    transform.node    false
elasticsearch-logstash-clients-0 100.96.11.192  100.96.11.192  ml.machine_memory 13958643712
elasticsearch-logstash-clients-0 100.96.11.192  100.96.11.192  ml.max_open_jobs  20
elasticsearch-logstash-clients-0 100.96.11.192  100.96.11.192  xpack.installed   true
elasticsearch-logstash-clients-0 100.96.11.192  100.96.11.192  transform.node    false
elasticsearch-data-2a-5          100.96.11.193  100.96.11.193  ml.machine_memory 122406567936
elasticsearch-data-2a-5          100.96.11.193  100.96.11.193  ml.max_open_jobs  20
elasticsearch-data-2a-5          100.96.11.193  100.96.11.193  xpack.installed   true
elasticsearch-data-2a-5          100.96.11.193  100.96.11.193  box_type          data
elasticsearch-data-2a-5          100.96.11.193  100.96.11.193  zone              2a
elasticsearch-data-2a-5          100.96.11.193  100.96.11.193  transform.node    true
elasticsearch-data-2a-4          100.96.0.5     100.96.0.5     ml.machine_memory 122406567936
elasticsearch-data-2a-4          100.96.0.5     100.96.0.5     ml.max_open_jobs  20
elasticsearch-data-2a-4          100.96.0.5     100.96.0.5     xpack.installed   true
elasticsearch-data-2a-4          100.96.0.5     100.96.0.5     box_type          data
elasticsearch-data-2a-4          100.96.0.5     100.96.0.5     zone              2a
elasticsearch-data-2a-4          100.96.0.5     100.96.0.5     transform.node    true
elasticsearch-data-2a-3          100.96.12.233  100.96.12.233  ml.machine_memory 122406567936
elasticsearch-data-2a-3          100.96.12.233  100.96.12.233  ml.max_open_jobs  20
elasticsearch-data-2a-3          100.96.12.233  100.96.12.233  xpack.installed   true
elasticsearch-data-2a-3          100.96.12.233  100.96.12.233  box_type          data
elasticsearch-data-2a-3          100.96.12.233  100.96.12.233  zone              2a
elasticsearch-data-2a-3          100.96.12.233  100.96.12.233  transform.node    true
elasticsearch-data-2a-2          100.96.38.96   100.96.38.96   ml.machine_memory 122406567936
elasticsearch-data-2a-2          100.96.38.96   100.96.38.96   ml.max_open_jobs  20
elasticsearch-data-2a-2          100.96.38.96   100.96.38.96   xpack.installed   true
elasticsearch-data-2a-2          100.96.38.96   100.96.38.96   box_type          data
elasticsearch-data-2a-2          100.96.38.96   100.96.38.96   zone              2a
elasticsearch-data-2a-2          100.96.38.96   100.96.38.96   transform.node    true
elasticsearch-data-2a-1          100.96.12.162  100.96.12.162  ml.machine_memory 122406567936
elasticsearch-data-2a-1          100.96.12.162  100.96.12.162  ml.max_open_jobs  20
elasticsearch-data-2a-1          100.96.12.162  100.96.12.162  xpack.installed   true
elasticsearch-data-2a-1          100.96.12.162  100.96.12.162  box_type          data
elasticsearch-data-2a-1          100.96.12.162  100.96.12.162  zone              2a
elasticsearch-data-2a-1          100.96.12.162  100.96.12.162  transform.node    true
elasticsearch-data-2a-0          100.96.20.3    100.96.20.3    ml.machine_memory 122406567936
elasticsearch-data-2a-0          100.96.20.3    100.96.20.3    ml.max_open_jobs  20
elasticsearch-data-2a-0          100.96.20.3    100.96.20.3    xpack.installed   true
elasticsearch-data-2a-0          100.96.20.3    100.96.20.3    box_type          data
elasticsearch-data-2a-0          100.96.20.3    100.96.20.3    zone              2a
elasticsearch-data-2a-0          100.96.20.3    100.96.20.3    transform.node    true
elasticsearch-client-0           100.96.1.132   100.96.1.132   ml.machine_memory 13958643712
elasticsearch-client-0           100.96.1.132   100.96.1.132   ml.max_open_jobs  20
elasticsearch-client-0           100.96.1.132   100.96.1.132   xpack.installed   true
elasticsearch-client-0           100.96.1.132   100.96.1.132   transform.node    false

Also yes, I have box_type attribute. I am using hot nodes only.

Can you change the box_type value from data to hot?

Okay, I changed that, and I see a different error now:

No full shingle in current detection window

Update: after keeping the detector in init state for a while, I got this error:

Aug 7, 2020 @ 14:49:43.799	Cannot get training data
Aug 7, 2020 @ 14:44:43.799	No RCF models are available either because RCF models are not ready or all nodes are unresponsive or the system might have bugs.
Aug 7, 2020 @ 14:39:43.799	No full shingle in current detection window

Do you continuous training data in your history? Without that, you will get “Cannot get training data”.