Docker Cookbook

276 | Chapter 9: Monitoring Containers

---

Restart your Docker daemon with 
sudo service docker restart
. You are now
ready to use 
curl
 and target the Docker remote API. The 
syntax
 is again simple: it is
an HTTP GET request to the 
/containers/(id)/stats
URI. Try it like so:
$ $ docker -H tcp://127.0.0.1:2375 run -d -p 5001:5000 runseb/flask
$ curl http://127.0.0.1:2375/containers/agitated_albattani/stats
{"read":"2015-04-01T11:48:40.609469913Z",\
"network":{"rx_bytes":648,"rx_packets":8,"...
Do not forget to replace 
agitated_albattani
with the name of your container. You
will start receiving a stream of statistics that you can interrupt with Ctrl-C. For practi‐
cal purposes, I truncated most of the results from the previous command. This is
handy for trying things out, but if (like me) you like Python, you might want to
access these statistics from a Python program. To do this, you can use 
docker-py
(see
Recipe 4.10
). A Python script like the one here will put you on the right track:
#!/usr/bin/env python
import
json
import
docker
import
sys
cli
=
docker
.
Client
(
base_url
=
'tcp://127.0.0.1:2375'
)
stats
=
cli
.
stats
(
sys
.
argv
[
1
])
print
json
.
dumps
(
json
.
loads
(
next
(
stats
)
.
rstrip
(
'\n'
)),
indent
=
4
)
The 
stats
object in this Python script is a 
generator
, which yields
results instead of the standard return behavior of functions. It is
used to capture the statistics stream and pick up where it left off.
next(stats)
in the script is the way to yield the latest result from
the stream.
See Also
• Original GitHub 
pull request
 for 
stats
• API 
documentation