Update list_hosts.py

examples/list_hosts.py

@@ -1,9 +1,16 @@
 #!/usr/bin/env python
 #
-# This script shows how to leverage sysdig's data query API to obtain the list
-# of the instrumented hosts that have been seen in the last 5 minutes.
+# This script shows how to leverage Sysdig data query API to obtain the list of the instrumented
+# hosts that have been seen in your infrastructure.
+# The output will show the container count (`container.count` metric) in addition to the
+# hostnames (`host.hostName` tag) in a format like this:
+#
+# host-1    12
+# host-2     4
+#
+# where the first column is the hostname and the second column is the number of containers running
+# in each host.
 #
-
 import getopt
 import json
 import os
@@ -15,62 +22,70 @@
 # Parse arguments
 #
 def usage():
-    print 'usage: %s [-j|--json] [-d|--duration <secs>] <sysdig-token>' % sys.argv[0]
-    print '-d|--duration: List hosts seen in the last <secs> seconds'
+    print 'usage: %s [-j|--json] [-d|--duration <secs>] [-c|--count <number>] <sysdig-token>' % sys.argv[0]
+    print '-d|--duration: List hosts seen in the last <secs> seconds (default: 3600, ie. last hour)'
+    print '-c|--count: Number of hosts to print (default: 100)'
     print '-j|--json: Print output as json'
     print 'You can find your token at https://app.sysdigcloud.com/#/settings/user'
     sys.exit(1)
-
 try:
-    opts, args = getopt.getopt(sys.argv[1:],"jd:",["json", "duration="])
+    opts, args = getopt.getopt(sys.argv[1:],"jd:c:",["json", "duration=", "count="])
 except getopt.GetoptError:
     usage()
 
-duration = 600
+duration = 3600
+count = 100
 print_json = False
 for opt, arg in opts:
     if opt in ("-d", "--duration"):
         duration = int(arg)
+    elif opt in ("-c", "--count"):
+        count = int(arg)
     elif opt in ("-j", "--json"):
         print_json = True
-
-if len(args) != 1:
-    usage()
-
 sdc_token = args[0]
 
-#
 # Instantiate the SDC client
-#
 sdclient = SdcClient(sdc_token)
 
 #
-# Prepare the query's metrics list. In this case, we have just one metric:
-# host.hostName. This is a 'key' metric, and we don't include any number metric.
-# Essentially, we create an 'enumeration' of hostnames.
+# Prepare the query's metrics list.
+# In this case, we have one tag (used for segmentation) and one metric:
+# - host.hostName. This is a tag, to identify each item of the output
+# - container.count: This is the metric
 #
-metrics = [{"id": "host.hostName"}]
+metrics = [
+    { "id": "host.hostName" },
+    { "id": "container.count", "aggregations": { "time": "avg", "group": "avg" } }
+]
 
-#
-# Fire the query.
-# Note: there's no sampling time. This means that we're requesting the result to
-#       come as a single sample.
-#
-res = sdclient.get_data(metrics, # metrics list
-                        -duration,   # cover the last duration seconds...
-                        0,      # ... ending now...
-                        duration)    # ... with just one durations sample
+res = sdclient.get_data(
+    metrics,  # list of metrics
+    -duration,  # start time: either a unix timestamp, or a difference from "now"
+    0,  # end time: either a unix timestamp, or a difference from "now" (0 means you need "last X seconds")
+    duration,  # sampling time, ie. data granularity;
+               # if equal to the time window span then the result will contain a single sample
+    paging = {
+        "from": 0,
+        "to": count - 1
+    })
 
-#
-# Show the results!
-#
 if res[0]:
-    data = res[1]
+    # data fetched successfully
+    if print_json:
+        print json.dumps(res[1])
+    else:
+        data = res[1]['data']
+        output = []
+        for i in range(0, len(data)):
+            sample = data[i]
+            metrics = sample['d']
+            hostName = metrics[0]
+            count = metrics[1]
+            output.append('%s\t%d' % (hostName, count))
+
+        print '\n'.join(output)
 else:
+    # data fetch failed
     print res[1]
     sys.exit(1)
-
-if print_json:
-    print json.dumps(data)
-else:
-    print data