docker daemon connection errors during tests killed test runs

While polling output on several test runs I noticed errors like the following:

stress-ng: info: [109717] stress-ng-af-alg: 19 cryptographic algorithms found in /proc/crypto
stress-ng: info: [109717] stress-ng-af-alg: 101 cryptographic algorithms max (with defconfigs)
stress-ng: info: [109729] stress-ng-stream: stressor loosely based on a variant of the STREAM benchmark code
stress-ng: info: [109729] stress-ng-stream: do NOT submit any of these results to the STREAM benchmark results
stress-ng: info: [109729] stress-ng-stream: Using CPU cache size of 5118K
time="2020-10-31T06:01:44Z" level=error msg="error waiting for container: unexpected EOF"
***********************************************

* Starting testflinger cleanup phase on saiph *

***********************************************

Cleaning up container if it exists...
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
complete

The following test nodes had these Docker daemon errors:

keylime bc4a4516-6e66-4d7f-a287-e2b19631574b
makrutlime 90fa28fd-7e0a-4c8e-88a4-4bda92ce8366
meitner 614a5b12-76f4-4c0a-8c1a-fe409d14fc6f
nabbit c7280a21-5425-49ee-9136-391f9ef8b962
persianlime b508b121-aa1e-4c24-8430-0f2164141412
prunes cc2b762a-db62-4aec-bd28-eb2a3be285c4
prunus dad02681-0d2b-4d8c-b918-5eb3a08c405f
saiph f5809f2e-647e-49bf-aebc-2c7e6c61164e

They all happened at different times and at different points in testing. Additionally, this killed the test and no results were uploaded.

So I'm guessing the docker containers died.