xingtanzjr commented on code in PR #9439:
URL: https://github.com/apache/iotdb/pull/9439#discussion_r1148902175
##########
server/src/main/java/org/apache/iotdb/db/service/metrics/IoTDBInternalLocalReporter.java:
##########
@@ -150,39 +150,43 @@ public ReporterType getReporterType() {
@Override
protected void writeMetricToIoTDB(Map<String, Object> valueMap, String
prefix, long time) {
- try {
- TSInsertRecordReq request = new TSInsertRecordReq();
- List<String> measurements = new ArrayList<>();
- List<TSDataType> types = new ArrayList<>();
- List<Object> values = new ArrayList<>();
- for (Map.Entry<String, Object> entry : valueMap.entrySet()) {
- String measurement = entry.getKey();
- Object value = entry.getValue();
- measurements.add(measurement);
- types.add(inferType(value));
- values.add(value);
- }
- ByteBuffer buffer = SessionUtils.getValueBuffer(types, values);
+ service.execute(
Review Comment:
Yes, you're correct. Actually my point of previous comment is that there is
`potential risk` in current implementation at the same time it won't lead to
dead-lock. This is just a talk about whether submit a new task inside a thread
pool to itself is a the best practice or not.
On the other hand, we can talk about another question, that is, we put the
Async logic inside the reporter's method rather than mark it in the flush logic
explicitly. If we have another implementation of Reporter, we cannot constraint
it with an ASYCN implementation, shall we add some limitation to it such as
change the method name to something like `xxxxxAsync` ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
