GabrielBrascher commented on a change in pull request #3078: Add influxdb to
statscollector
URL: https://github.com/apache/cloudstack/pull/3078#discussion_r246011524
##########
File path: server/src/main/java/com/cloud/server/StatsCollector.java
##########
@@ -497,86 +572,35 @@ protected void runInContext() {
}
try {
- HashMap<Long, VmStatsEntry> vmStatsById =
_userVmMgr.getVirtualMachineStatistics(host.getId(), host.getName(), vmIds);
+ Map<Long, VmStatsEntry> vmStatsById =
_userVmMgr.getVirtualMachineStatistics(host.getId(), host.getName(), vmIds);
if (vmStatsById != null) {
- VmStatsEntry statsInMemory = null;
-
Set<Long> vmIdSet = vmStatsById.keySet();
for (Long vmId : vmIdSet) {
VmStatsEntry statsForCurrentIteration =
vmStatsById.get(vmId);
- statsInMemory =
(VmStatsEntry)_VmStats.get(vmId);
+ statsForCurrentIteration.setVmId(vmId);
+ UserVmVO userVmVo = _userVmDao.findById(vmId);
+ statsForCurrentIteration.setUserVmVO(userVmVo);
- if (statsInMemory == null) {
- //no stats exist for this vm, directly
persist
- _VmStats.put(vmId,
statsForCurrentIteration);
- } else {
- //update each field
-
statsInMemory.setCPUUtilization(statsForCurrentIteration.getCPUUtilization());
-
statsInMemory.setNumCPUs(statsForCurrentIteration.getNumCPUs());
-
statsInMemory.setNetworkReadKBs(statsInMemory.getNetworkReadKBs() +
statsForCurrentIteration.getNetworkReadKBs());
-
statsInMemory.setNetworkWriteKBs(statsInMemory.getNetworkWriteKBs() +
statsForCurrentIteration.getNetworkWriteKBs());
-
statsInMemory.setDiskWriteKBs(statsInMemory.getDiskWriteKBs() +
statsForCurrentIteration.getDiskWriteKBs());
-
statsInMemory.setDiskReadIOs(statsInMemory.getDiskReadIOs() +
statsForCurrentIteration.getDiskReadIOs());
-
statsInMemory.setDiskWriteIOs(statsInMemory.getDiskWriteIOs() +
statsForCurrentIteration.getDiskWriteIOs());
-
statsInMemory.setDiskReadKBs(statsInMemory.getDiskReadKBs() +
statsForCurrentIteration.getDiskReadKBs());
-
statsInMemory.setMemoryKBs(statsForCurrentIteration.getMemoryKBs());
-
statsInMemory.setIntFreeMemoryKBs(statsForCurrentIteration.getIntFreeMemoryKBs());
-
statsInMemory.setTargetMemoryKBs(statsForCurrentIteration.getTargetMemoryKBs());
-
- _VmStats.put(vmId, statsInMemory);
- }
-
- /**
- * Add statistics to HashMap only when they
should be send to a external stats collector
- * Performance wise it seems best to only
append to the HashMap when needed
- */
- if (externalStatsEnabled) {
- VMInstanceVO vmVO =
_vmInstance.findById(vmId);
- String vmName = vmVO.getUuid();
-
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".cpu.num",
statsForCurrentIteration.getNumCPUs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".cpu.utilization",
statsForCurrentIteration.getCPUUtilization());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".network.read_kbs",
statsForCurrentIteration.getNetworkReadKBs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".network.write_kbs",
statsForCurrentIteration.getNetworkWriteKBs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".disk.write_kbs",
statsForCurrentIteration.getDiskWriteKBs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".disk.read_kbs",
statsForCurrentIteration.getDiskReadKBs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".disk.write_iops",
statsForCurrentIteration.getDiskWriteIOs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".disk.read_iops",
statsForCurrentIteration.getDiskReadIOs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".memory.total_kbs",
statsForCurrentIteration.getMemoryKBs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".memory.internalfree_kbs",
statsForCurrentIteration.getIntFreeMemoryKBs());
- metrics.put(externalStatsPrefix +
"cloudstack.stats.instances." + vmName + ".memory.target_kbs",
statsForCurrentIteration.getTargetMemoryKBs());
+
storeVirtualMachineStatsInMemory(statsForCurrentIteration);
+ if (externalStatsType ==
ExternalStatsProtocol.GRAPHITE) {
+ prepareVmMetricsForGraphite(metrics,
statsForCurrentIteration);
+ } else {
+
metrics.put(statsForCurrentIteration.getVmId(), statsForCurrentIteration);
}
-
}
- /**
- * Send the metrics to a external stats collector
- * We send it on a per-host basis to prevent that
we flood the host
- * Currently only Graphite is supported
- */
if (!metrics.isEmpty()) {
- if (externalStatsType != null &&
externalStatsType == ExternalStatsProtocol.GRAPHITE) {
-
- if (externalStatsPort == -1) {
- externalStatsPort = 2003;
- }
-
- s_logger.debug("Sending VmStats of host "
+ host.getId() + " to Graphite host " + externalStatsHost + ":" +
externalStatsPort);
-
- try {
- GraphiteClient g = new
GraphiteClient(externalStatsHost, externalStatsPort);
- g.sendMetrics(metrics);
- } catch (GraphiteException e) {
- s_logger.debug("Failed sending VmStats
to Graphite host " + externalStatsHost + ":" + externalStatsPort + ": " +
e.getMessage());
- }
-
- metrics.clear();
+ if (externalStatsType ==
ExternalStatsProtocol.GRAPHITE) {
+ sendVmMetricsToGraphiteHost(metrics, host);
+ } else if (externalStatsType ==
ExternalStatsProtocol.INFLUXDB) {
+ sendMetricsToInfluxdb(metrics);
Review comment:
@DaanHoogland sorry, the lines 600 - 603 have changed, now are 596 and 598.
I was referring to the following mthods
`sendVmMetricsToGraphiteHost(metrics, host);` and
`sendMetricsToInfluxdb(metrics);`
I had two options: (i) keep the Graphite metrics Map as it was, and create a
new map for the influxdb (for instance metricsInfluxDB):
```
if (externalStatsType == ExternalStatsProtocol.GRAPHITE) {
sendVmMetricsToGraphiteHost(metrics, host);
} else if (externalStatsType == ExternalStatsProtocol.INFLUXDB) {
sendMetricsToInfluxdb(metricsInfluxDB);
```
Or (ii) change the metrics map from `Map<String, Integer>` to a generic
`Map<Object, Object>`, which implicates on changing the GraphiteClient.java
sendMetrics method, and reuse the `metrics` for InfluxDB and Graphite.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
