def query_cluster_status(request): #统计每个集群的可用性
try:
cluster = request.POST['cluster']
range = request.POST['range']
end_timestamp = int(time.time())
start_timestamp = end_timestamp - int(range)
#先统计有哪些集群名
if cluster == 'all':
cluster_list = [i['fcluster'] for i in Service.objects.values('fcluster').distinct() if i['fcluster']]
else:
cluster_list = [cluster]
counters = ['cluster.available.percent/clusterName=%s,project=oms'%i for i in cluster_list]
endpoint = get_local_ip()
f = Falcon()
history_data = f.get_history_data(start_timestamp, end_timestamp, [endpoint], counters, step=24*60*60, CF='AVERAGE')
data = []
for i in history_data:
if i['counter']:
tags = i['counter'].split('/')[1]
tag_dict = gen_tags(tags)
cluster = tag_dict.get('clusterName')
ts_value = []
c_date = time.strftime("%Y-%m-%d")
c_timestamp = int(time.mktime(time.strptime(c_date, '%Y-%m-%d'))) * 1000
for j in i['Values']:
if (j['timestamp'] - 8*60*60)*1000 == c_timestamp:
ts_value.append([c_timestamp, get_cluster_available(cluster, c_date)])
else:
ts_value.append([(j['timestamp'] - 8*60*60)*1000, j['value']]) #这里要减去8个小时，是因为rrd里存的点的时刻是8点钟
data.append({'data':ts_value, 'name':cluster})
except:
print traceback.format_exc()
return JsonResponse({'code': 0, 'data': data, 'message': 'ok'})
def query_graph(self, request):
fid = request.POST['fid']
range = request.POST['range']
end_ts = int(time.time())
start_ts = end_ts - int(range)
service_obj = Service.objects.get(fid=fid)
endpoint = service_obj.fhost
#统计该组件对应的大metric有哪些
all_data = []
metric_list = []
f = Falcon()
for metric, item_list in items[service_obj.fname].items():
metric_list.append(metric)
counter_list = ['%s/%sport=%s'%(i, service_obj.fname, service_obj.fport) for i in item_list]
#print counter_list
history_data = f.get_history_data(start_ts, end_ts, [endpoint], counter_list, CF='AVERAGE')
#print history_data[0]
hdata = []
for i in history_data:
hdata.append({'name':i['counter'], 'data':[[j['timestamp']*1000, j['value'] ]for j in i['Values']]})
all_data.append({'metric':metric, 'hdata':hdata})
print all_data
#同时渲染文件
f = open('templates/service/chart_div_tmp.html')
t = template.Template(f.read())
con = template.Context({'data': serialize_number(len(all_data), 2), 'metric_list':metric_list})
res = t.render(con)
f.close()
return JsonResponse({'code':0, 'data':{'all_data':all_data, 'html_content':res}, 'message':'ok'})
def query_port(self, request):
host = request.POST['host']
port = request.POST['port']
start_time = request.POST['start_time']
end_time = request.POST['end_time']
start_ts = int(time.mktime(time.strptime(start_time,'%Y-%m-%d %H:%M:%S')))
end_ts = int(time.mktime(time.strptime(end_time, '%Y-%m-%d %H:%M:%S')))
counter = '%s/port=%s,project=oms'%(settings.port_listen_key, port)
f = Falcon()
history_data = f.get_history_data(start_ts, end_ts, [host], [counter], CF='AVERAGE')
hdata = []
for i in history_data:
hdata.append({'name': i['endpoint'], 'data': [[j['timestamp'] * 1000, j['value']] for j in i['Values']]})
return JsonResponse({'code': 0, 'data': {'hdata': hdata}, 'message': 'ok'})
def query_service_top(self, request):
try:
cluster = request.POST['cluster']
date = request.POST['date']
topn = int(request.POST['topn'])
start_ts = int(time.mktime(time.strptime(date, '%Y-%m-%d')))
end_ts = start_ts + 86400
#先插集群下有哪些ip和端口
endpoints = []
ports = []
for i in Service.objects.filter(fcluster=cluster):
if i.fhost:
endpoints.append(i.fhost)
if i.fport:
ports.append(i.fport)
endpoints = list(set(endpoints))
ports = list(set(ports))
counters = ['%s/port=%s,project=oms' % (settings.port_listen_key, i) for i in ports]
f = Falcon()
history_data = f.get_history_data(start_ts, end_ts, endpoints, counters, step=60, CF='AVERAGE')
print history_data
names = []
values = []
data = {}
for i in history_data:
if i['Values']:
host = i['endpoint']
tags = i['counter'].split('/')[1]
tag_dict = gen_tags(tags)
port = tag_dict['port']
if not Service.objects.filter(fhost=host,fport=port,fcluster=cluster):
continue
# 计算故障率
fail_count = 0
for j in i['Values']:
if j['value'] == 0:
fail_count += 1
fail_rate = '%.2f' % (float(fail_count) / len(i['Values']) * 100)
fail_rate = float(fail_rate)
data['%s:%s' % (host, port)] = fail_rate
# 排序
if data:
print data
data = sorted(data.items(), key=lambda x: x[1], reverse=True)[:topn]
for k, v in data:
names.append(k)
values.append(v)
except:
print traceback.format_exc()
return JsonResponse({'code': 0, 'data': {'names': names, 'values': values}, 'message': 'ok'})
def query_graph2(self, request):
fid = request.POST['fid']
counter = request.POST['counter']
range = request.POST['range']
end_ts = int(time.time())
start_ts = end_ts - int(range)
service_obj = Service.objects.get(fid=fid)
endpoint = service_obj.fhost
f = Falcon()
history_data = f.get_history_data(start_ts, end_ts, [endpoint], [counter], CF='AVERAGE')
# print history_data[0]
hdata = []
for i in history_data:
hdata.append({'name': i['endpoint'], 'data': [[j['timestamp']*1000, j['value']] for j in i['Values']]})
return JsonResponse({'code': 0, 'data': {'hdata': hdata}, 'message': 'ok'})
def get_cluster_available(cluster, c_date): #获取某一天某个集群的可用率
start_ts = int(time.mktime(time.strptime(c_date, '%Y-%m-%d')))
end_ts = start_ts + 86400
host_list = []
port_list = []
weight_info = {}
for j in Service.objects.filter(fcluster=cluster):
if j.fhost:
host_list.append(j.fhost)
if j.fport:
port_list.append(str(j.fport))
if j.fhost and j.fport:
weight_info['%s_%s' % (j.fhost, j.fport)] = j.fweight
host_list = list(set(host_list))
port_list = list(set(port_list))
counter_list = ['listen.port/port=%s,project=oms' % i for i in port_list]
f = Falcon()
data = f.get_history_data(start_ts,
end_ts,
host_list,
counter_list,
step=60)
#print data
success_count = 0
total = 0
for i in data:
values = i['Values']
counter = i['counter']
endpoint = i['endpoint']
tags = counter.split('/')[1]
tag_dict = gen_tags(tags)
port = tag_dict['port']
if values:
weight = weight_info.get('%s_%s' % (endpoint, port), 1)
success_list = [j['value'] for j in values if j['value'] != 0]
success_count += weight * len(success_list)
total += weight * len(values)
if success_count == 0:
cluster_available_rate = 100
else:
cluster_available_rate = float('%.2f' %
(float(success_count) / total * 100))
return cluster_available_rate
def query_service_top(request):
range = request.POST['range']
topn = int(request.POST['topn'])
end_timestamp = int(time.time())
start_timestamp = end_timestamp - int(range)
#先统计有endporints和couters
endpoints = []
ports = []
for i in Service.objects.all():
if i.fhost:
endpoints.append(i.fhost)
if i.fport:
ports.append(i.fport)
endpoints = list(set(endpoints))
ports = list(set(ports))
counters = ['%s/port=%s,project=oms'%(settings.port_listen_key, i) for i in ports]
f = Falcon()
history_data = f.get_history_data(start_timestamp, end_timestamp, endpoints, counters, step=60, CF='AVERAGE')
names = []
values = []
data = {}
for i in history_data:
if i['Values']:
host = i['endpoint']
tags = i['counter'].split('/')[1]
tag_dict = gen_tags(tags)
port = tag_dict['port']
#计算故障率
fail_count = 0
for j in i['Values']:
if j['value'] == 0:
fail_count += 1
fail_rate = '%.2f'%(float(fail_count)/len(i['Values'])*100)
fail_rate = float(fail_rate)
data['%s:%s'%(host, port)] = fail_rate
#排序
data = sorted(data.items(), key=lambda x: x[1], reverse=True)[:topn]
for k, v in data:
names.append(k)
values.append(v)
print values
return JsonResponse({'code': 0, 'data': {'names':names, 'values':values}, 'message': 'ok'})