README.rst

Design Notes

Databases:

• docker-beta_cloud_dst_ibm_com:8080
• rancher_index

Collections inside rancher_index:

• endpoints

Collections inside docker-beta_cloud_dst_ibm_com:8080:

• collection_dates
• containers
• dockerhosts
• environments
• last_ts
• services
• stacks

last_ts is a single key/value pair that has the ts timestamp of the last collection date completed. This ts value is then used to get the relevant latest collection data. The value is Unix epoch seconds. Therefore, calculate accordingly from client side, to determine if too old.

To narrow the return fields, do it like this when you call find():

from pymongo import MongoClient
from pprint import pprint as pp

cli = MongoClient(host = 'mopbz171199.fr.dst.ibm.com')
DB = cli.get_database('docker-beta_cloud_dst_ibm_com:8080')
latest = DB.last_ts.find().next()

filter1 = {'ts':latest['value']}
filter2 = {}
filter3 = {'ts':1470862817.261788}

containers = DB.containers.find(filter1,
    {'name':1,'state':1,'containerHost':1})

for container in containers:
    #pp(container)
    print container['name'],container['state'],container['containerHost']
    print "-" * 20

The first parameter of find() is a filter dictionary. If it is empty, it means get all the documents (aka rows) from the collection (aka table). The second parameter of find() is used to narrow the fields of interest. From the snippet above, only shows interest to the fields: name, state and containerHost.

Filter meanings

• filter1 = get latest containers data for this rancher host
• filter2 = get all containers data for this rancher host
• filter3 = get containers data for the specified timestamp

Test run url: http://smartcast.boulder.ibm.com:81?session=8d5f36ab-7f4a-4130-aafe-688a7bd44a87