unauthorized access error from orbit-datasource plugin

[nilivingston]

While setting up cosmos-web, I discovered that in Grafana I needed to go to Configuration -> Plugins and select the orbit-datasource plugin, then click Create a orbit-datasource data source. This caused orbit-datasource to appear in Grafana's Data sources list, and I was able to use the Cosmos Simulator panel to create some data in InfluxDB.

I updated the Orbit Display panel to use orbit-datasource and set the Name as node0 and checked the Simulation Mode toggle. After saving the panel, I receive a warning icon in Grafana with the message unauthorized: unauthorized access.

The error message from the InfluxDB logs is:
influxdb_1 | ts=2022-08-08T23:13:49.428872Z lvl=info msg=Unauthorized log_id=0cBtjPH0000 error="authorization not found"

And the related log entry from Grafana is:
grafana_1 | logger=plugin.myorg-orbit-datasource t=2022-08-08T23:13:49.428139216Z level=info msg="QueryData called" request="map[Headers:<nil> PluginContext:map[AppInstanceSettings:<nil> DataSourceInstanceSettings:map[BasicAuthEnabled:false BasicAuthUser: Database: DecryptedSecureJSONData:map[apiKey:REDACTED_MY_API_KEY] ID:4 JSONData:map[] Name:orbit-datasource UID:qnPO1mmVz URL: Updated:2022-08-08T23:03:10Z User:] OrgID:1 PluginID:myorg-orbit-datasource User:map[Email:admin@localhost Login:admin Name: Role:Admin]] Queries:[map[Interval:2e+10 JSON:map[constant:6.5 datasource:map[type:myorg-orbit-datasource uid:qnPO1mmVz] datasourceId:4 intervalMs:20000 isSimMode:true maxDataPoints:1044 opNodeList:[map[name: px:node.loc.pos.eci.s.col[0] py:node.loc.pos.eci.s.col[1] pz:node.loc.pos.eci.s.col[2] tag_name:beacon_type tag_value:posbeacon vx:node.loc.pos.eci.v.col[0] vy:node.loc.pos.eci.v.col[1] vz:node.loc.pos.eci.v.col[2]]] refId:A simNodeList:[map[frame:eci name:node0]]] MaxDataPoints:1044 QueryType: RefID:A TimeRange:map[From:2022-08-08T17:13:49.415Z To:2022-08-08T23:13:49.415Z]]]]"

I updated the plugin settings for orbit-datasource with a valid API Key and received the same error. Testing the datasource using the Save & test button works, but using the Explore option produces the same error. I have not added any value for the Path setting in the orbit-datasource configuration.

[TylerGraham74]

Hi, I am having the same issue. Has this been resolved or have you figured out a work around?

[nilivingston]

Hi Tyler. I was not able to come up with a workaround. I think it should be possible to fork the plugin and adjust the authentication. I have not had time to give it a try though, sorry.

[ginoscott]

See if the latest commits (commit1 commit 2) addresses these issues, which have removed the hardcoded host url, authentication key, and organization into settings that can be configured from the datasource settings.

[TylerGraham74]

That looks like it fixed the described error, still having problems with the orbit display. Now, I get an error saying "Query returned no results." Any idea on what could be causing this now?

[nilivingston]

I updated my Grafana deployment to use the new version of the orbit-datasource plugin. I'm now getting an error in the dashboard that says "Error loading: hsfl-orbit-display". The Grafana logs show:

logger=plugin.hsfl-orbit-datasource t=2022-10-25T17:55:26.991214384Z level=error msg="query error1" EXTRA_VALUE_AT_END="invalid: cannot access element 0 of array of length 0"
logger=plugin.hsfl-orbit-datasource t=2022-10-25T17:55:42.903873647Z level=error msg="Host to IP error" EXTRA_VALUE_AT_END="lookup cosmos_web_backend on 10.100.0.10:53: no such host"

I'm not seeing any of the Cesium visuals come up in the dashboard panels either, so I'm wondering if that's part of the issue.

[ginoscott]

Query returned no results."

I've left an empty query result as an error, that's probably confusing, I'll get rid of it. It also means that nothing was sent to the influxdb database from the cosmos sim panel (which sends some simulated orbit data), there is no data that was retrieved.

There are a few pieces to getting the whole thing strung together. It starts with the cosmos sim panel being able to make the backend call properly:

const ds = await getDataSourceSrv().get('orbit-datasource');
const observable = getBackendSrv()
        .fetch<CallResourcesResponse>({
          method: 'POST',
          url: `api/datasources/${ds.id}/resources/propagator_db`,
          data: payload,
        })
        .pipe(map((response) => response));

Orbit datasource has the endpoint, where it does another api call to a nodejs server running in the cosmos_web_backend container to run the orbit data generator. So the first thing I should confirm is if:

logger=plugin.hsfl-orbit-datasource t=2022-10-25T17:55:42.903873647Z level=error msg="Host to IP error" EXTRA_VALUE_AT_END="lookup cosmos_web_backend on 10.100.0.10:53: no such host"

the cosmos_web_backend container is running?

And I also want to clarify if by:

I'm not seeing any of the Cesium visuals come up in the dashboard panels either, so I'm wondering if that's part of the issue.

Do you mean that the orbit-display panel is showing a blank panel with nothing in it?

[nilivingston]

Yesterday when I created the panel it was blank, although when I fired it up this morning to check I see the nice Cesium globe now:

Looks like for that part I just needed to be a bit more patient.

I tried the "submit" button on the simulation panel and these are the log results from about that time:

logger=plugin.hsfl-orbit-datasource t=2022-10-26T13:42:02.153522106Z level=info msg=queryModel qm="map[isSimMode:true opNodeList:[map[name: px:node.loc.pos.eci.s.col[0] py:node.loc.pos.eci.s.col[1] pz:node.loc.pos.eci.s.col[2] tag_name:beacon_type tag_value:posbeacon vx:node.loc.pos.eci.v.col[0] vy:node.loc.pos.eci.v.col[1] vz:node.loc.pos.eci.v.col[2]]] simNodeList:[map[frame:eci name:node0]]]"
logger=plugin.hsfl-orbit-datasource t=2022-10-26T13:42:02.323193767Z level=error msg="query error1" EXTRA_VALUE_AT_END="invalid: cannot access element 0 of array of length 0"
logger=plugin.hsfl-orbit-datasource t=2022-10-26T13:42:22.399785895Z level=error msg="Host to IP error" EXTRA_VALUE_AT_END="lookup cosmos_web_backend on 10.100.0.10:53: no such host"`

[ginoscott]

logger=plugin.hsfl-orbit-datasource t=2022-10-26T13:42:22.399785895Z level=error msg="Host to IP error" EXTRA_VALUE_AT_END="lookup cosmos_web_backend on 10.100.0.10:53: no such host"

So this is the bit that's trying to access cosmos_web_backend, which docker-compose automatically creates an alias for containers in the same network to access by. Are you running your setup with the docker-compose up -d? If the cosmos_web_backend container is running and you're seeing this message, then that's an issue.

[nilivingston]

I'm running everything on a Kubernetes cluster. I created a Docker image and a Helm chart for the Cosmos backend and I have that deployed. Now that you mention it, I haven't configured anything in Grafana that would point to the Cosmos backend service so it makes sense that the plugin can't reach it. Is there an option somewhere for configuring the backend value so I can change it to the service address?

[farrell-j]

I'm running everything on a Kubernetes cluster. I created a Docker image and a Helm chart for the Cosmos backend and I have that deployed. Now that you mention it, I haven't configured anything in Grafana that would point to the Cosmos backend service so it makes sense that the plugin can't reach it. Is there an option somewhere for configuring the backend value so I can change it to the service address?

Did you ever find a solution to this issue?

[nblivin]

I'm running everything on a Kubernetes cluster. I created a Docker image and a Helm chart for the Cosmos backend and I have that deployed. Now that you mention it, I haven't configured anything in Grafana that would point to the Cosmos backend service so it makes sense that the plugin can't reach it. Is there an option somewhere for configuring the backend value so I can change it to the service address?

Did you ever find a solution to this issue?

I did not. I ended up switching to OpenC3 instead.