)]}'
{"monasca_agent/collector/checks_d/ovs.py":[{"author":{"_account_id":15130,"name":"Sonu","email":"sonu.sudhakaran@gmail.com","username":"sonu"},"change_message_id":"8521db5a30862b1cfceb7d737978923aa2870c72","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"7a3c09a3_5fbcf805","updated":"2017-01-17 14:21:32.000000000","message":"Do you get the max bandwidth from OVS? What are the means of calculating the capacity? \n\nI tried reading though the lines, but got confused when saw a mention of novaclient.","commit_id":"d038bd27835377273567fad7f52ad49ca5601d29"},{"author":{"_account_id":8126,"name":"Bradley Klein","email":"bradley.klein@charter.com","username":"bklei"},"change_message_id":"5da205b790d00458e48a0bcb5e3035d622458d42","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"5a3905b3_000454be","in_reply_to":"7a3c09a3_5fbcf805","updated":"2017-01-19 15:24:28.000000000","message":"until neutron supports QOS at the router level, we can calculate based on the bandwidth limits set in nova (if set, that is) for instances using the router.  that\u0027s why we need to make the nova calls, see https://wiki.openstack.org/wiki/InstanceResourceQuota#Bandwidth_limits","commit_id":"d038bd27835377273567fad7f52ad49ca5601d29"},{"author":{"_account_id":15130,"name":"Sonu","email":"sonu.sudhakaran@gmail.com","username":"sonu"},"change_message_id":"8521db5a30862b1cfceb7d737978923aa2870c72","unresolved":false,"context_lines":[{"line_number":502,"context_line":"        burst_bw \u003d 0"},{"line_number":503,"context_line":""},{"line_number":504,"context_line":"        #"},{"line_number":505,"context_line":"        # we\u0027ll sum inbound and outbound for the max possible throughput"},{"line_number":506,"context_line":"        #"},{"line_number":507,"context_line":"        avg_re \u003d re.compile(\u0027quota:vif_.*bound_average\u0027)"},{"line_number":508,"context_line":"        peak_re \u003d re.compile(\u0027quota:vif_.*bound_peak\u0027)"}],"source_content_type":"text/x-python","patch_set":3,"id":"7a3c09a3_5f6078cf","line":505,"range":{"start_line":505,"start_character":53,"end_line":505,"end_character":61},"updated":"2017-01-17 14:21:32.000000000","message":"I am sure, you mean capacity and maximum throughput differently. Since they mean different since throughput is limited by performance.","commit_id":"d038bd27835377273567fad7f52ad49ca5601d29"},{"author":{"_account_id":8126,"name":"Bradley Klein","email":"bradley.klein@charter.com","username":"bklei"},"change_message_id":"5da205b790d00458e48a0bcb5e3035d622458d42","unresolved":false,"context_lines":[{"line_number":502,"context_line":"        burst_bw \u003d 0"},{"line_number":503,"context_line":""},{"line_number":504,"context_line":"        #"},{"line_number":505,"context_line":"        # we\u0027ll sum inbound and outbound for the max possible throughput"},{"line_number":506,"context_line":"        #"},{"line_number":507,"context_line":"        avg_re \u003d re.compile(\u0027quota:vif_.*bound_average\u0027)"},{"line_number":508,"context_line":"        peak_re \u003d re.compile(\u0027quota:vif_.*bound_peak\u0027)"}],"source_content_type":"text/x-python","patch_set":3,"id":"5a3905b3_6030b857","line":505,"range":{"start_line":505,"start_character":53,"end_line":505,"end_character":61},"in_reply_to":"7a3c09a3_5f6078cf","updated":"2017-01-19 15:24:28.000000000","message":"just trying to publish what the maximum possible traffic through the router can be based on quotas set in instances using the router.  let me know how better to describe that if you have ideas.","commit_id":"d038bd27835377273567fad7f52ad49ca5601d29"},{"author":{"_account_id":15130,"name":"Sonu","email":"sonu.sudhakaran@gmail.com","username":"sonu"},"change_message_id":"8521db5a30862b1cfceb7d737978923aa2870c72","unresolved":false,"context_lines":[{"line_number":518,"context_line":""},{"line_number":519,"context_line":"        return max(avg_bw, peak_bw, burst_bw)"},{"line_number":520,"context_line":""},{"line_number":521,"context_line":"    def _add_max_bw_to_port_cache(self, port_cache, all_ports_data):"},{"line_number":522,"context_line":"        if not self.publish_router_capacity:"},{"line_number":523,"context_line":"            return port_cache"},{"line_number":524,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"7a3c09a3_0c3eccd9","line":521,"range":{"start_line":521,"start_character":0,"end_line":521,"end_character":4},"updated":"2017-01-17 14:21:32.000000000","message":"I am unable to gauge why we need to loop through all instances in port_cache, when we have a loop to enumerate all instances using OVS directly, and get those instances from cache directly.\n\nLooping over a cache is not a good idea since they may give inconsistent results as the state of caches can never be guaranteed.\n\nWhy can\u0027t we integrate the max b/w computation in the loop starting at #138?","commit_id":"d038bd27835377273567fad7f52ad49ca5601d29"},{"author":{"_account_id":8126,"name":"Bradley Klein","email":"bradley.klein@charter.com","username":"bklei"},"change_message_id":"5da205b790d00458e48a0bcb5e3035d622458d42","unresolved":false,"context_lines":[{"line_number":518,"context_line":""},{"line_number":519,"context_line":"        return max(avg_bw, peak_bw, burst_bw)"},{"line_number":520,"context_line":""},{"line_number":521,"context_line":"    def _add_max_bw_to_port_cache(self, port_cache, all_ports_data):"},{"line_number":522,"context_line":"        if not self.publish_router_capacity:"},{"line_number":523,"context_line":"            return port_cache"},{"line_number":524,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"5a3905b3_e0f58883","line":521,"range":{"start_line":521,"start_character":0,"end_line":521,"end_character":4},"in_reply_to":"7a3c09a3_0c3eccd9","updated":"2017-01-19 15:24:28.000000000","message":"i\u0027m being intentional not to make additional ovs calls, we update ovs ports once per cache refresh interval (defaults to 4 hours), and the existing cache (that pre-dates this patch) contains all the data necessary to determine which instances are using the router in question.","commit_id":"d038bd27835377273567fad7f52ad49ca5601d29"}]}