Inspect failed and running tasks of all services

task_inspect_capture.sh

## Diagnostic data capture for Docker services and their associated tasks
## Run where client bundle is loaded or on a swarm manager 
## Use something like: 
## docker service ps $each | grep -v 'ID' | awk '{print $1 " " $6}' | egrep -i 'failed|running' | awk '{print $1}';
## to grep out specific states within TASK_ID
## ---

# Create a directory to store this data in
DIRECTORY="$(pwd)/docker-service-info/"
mkdir -p $DIRECTORY

# Capture node, service and stack level information
echo "Capturing node, service and stack level information..."
docker node ls > $DIRECTORY/docker-node-ls.out
docker stack ls > $DIRECTORY/docker-stack-ls.out
docker service ls > $DIRECTORY/docker-service-ls.out

# Iterate through each of the running services and collect info
echo "Starting service info capture..."
SERVICES=$(docker service ls -q)

for each in $SERVICES
do
    SERVICENAME=$(docker inspect $each --format {{.Spec.Name}})
    mkdir -p $DIRECTORY/$SERVICENAME
    docker service ps $each > $DIRECTORY/$SERVICENAME/docker-ps.out
    docker service inspect $each > $DIRECTORY/$SERVICENAME/docker-service-inspect.out
    # Back to awk: Provide backwards compat for engines which lack --format
    for TASK_ID in $(
        docker service ps $each | grep -v 'ID' | awk '{print $1}'; 
    )                
    do
         docker inspect --type=task $TASK_ID > $DIRECTORY/$SERVICENAME/$TASK_ID-task-inspect.out
    done
    echo $SERVICENAME" captured"
done

# Create a tarball of the directory contents, if tar exists
if [ -e /bin/tar ]; then
    echo "Creating docker-service-info.tar.gz of captured information..."
    tar czf docker-service-info.tar.gz $DIRECTORY
    echo "Done."
fi