grafana postgres data source test returns "Metric request error" - postgresql

Question

I have deployed a docker container using the TimescaleDB extension to Postgres as a persistent store for metrics collected by Prometheus. This includes deploying the Prometheus Postgres adapter in a separate container. And then I deployed grafana in a third container.
Inside the TimescaleDB container I can execute the query:
SELECT time_bucket('5 minutes', time) AS five_min_bucket, name, avg(value)
FROM metrics
WHERE (name='node_load5' OR name='node_memory_Active_bytes') AND
time > NOW() - interval '1 day'
GROUP BY five_min_bucket,name
ORDER BY five_min_bucket;
and it returns these rows from the metrics table:
five_min_bucket | name | avg
------------------------+--------------------------+-------------------
2019-07-13 16:15:00+00 | node_load5 | 1.84263157894737
2019-07-13 16:15:00+00 | node_memory_Active_bytes | 1055776552.42105
2019-07-13 16:20:00+00 | node_load5 | 1.62533333333333
2019-07-13 16:20:00+00 | node_memory_Active_bytes | 1070225817.6
2019-07-13 16:25:00+00 | node_load5 | 1.06166666666667
2019-07-13 16:25:00+00 | node_memory_Active_bytes | 1084650018.13333
2019-07-13 16:30:00+00 | node_load5 | 0.593333333333333
2019-07-13 16:30:00+00 | node_memory_Active_bytes | 1093145668.26667
2019-07-13 16:35:00+00 | node_load5 | 0.322333333333333
2019-07-13 16:35:00+00 | node_memory_Active_bytes | 1108816145.06667
2019-07-13 16:40:00+00 | node_load5 | 0.180666666666667
2019-07-13 16:40:00+00 | node_memory_Active_bytes | 1125213388.8
2019-07-13 16:45:00+00 | node_load5 | 0.126
2019-07-13 16:45:00+00 | node_memory_Active_bytes | 1134488098.13333
2019-07-13 16:50:00+00 | node_load5 | 0.133333333333333
2019-07-13 16:50:00+00 | node_memory_Active_bytes | 1162440704
(16 rows)
The grafana container is run as:
docker run \
-d \
-p 3000:3000 \
--name=grafana \
-v grafana-storage:/var/lib/grafana \
-e "GF_INSTALL_PLUGINS=grafana-kubernetes-app" \
grafana/grafana
When I launch a browser to http://104.XX.XX.XX:3000 I can log into the web-ui and define the Datasource to postgres, but I received the error 'Metric request error'.
I looked for the grafana logs in its container and the path /var/log/grafana is empty. This is the log location indicated when I do a docker inspect command.
Using Chrome, I am able to inspect the Console and it reports the server responded with an internal error of 500.
On the host, I can confirm the docker-proxy is listening on port 5432.
sudo netstat -antp | grep "5432"
tcp 0 0 172.17.0.1:49486 172.17.0.2:5432 ESTABLISHED 60631/docker-proxy
tcp6 0 0 :::5432 :::* LISTEN 60631/docker-proxy
tcp6 0 0 10.0.0.4:5432 172.18.0.2:38568 ESTABLISHED 60631/docker-proxy
How do I get grafana to fetch metrics stored in the postgres database?