Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/4688#discussion_r140291389
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/rest/handler/legacy/CurrentJobsOverviewHandler.java
---
@@ -92,24 +101,26 @@ public CurrentJobsOverviewHandler(
StringWriter writer = new
StringWriter();
try {
JsonGenerator gen =
JsonFactory.JACKSON_FACTORY.createGenerator(writer);
+ final
JobDetails.JobDetailsSerializer jobDetailsSerializer = new
JobDetails.JobDetailsSerializer();
+
gen.writeStartObject();
if (includeRunningJobs &&
includeFinishedJobs) {
gen.writeArrayFieldStart("running");
- for (JobDetails detail
: result.getRunningJobs()) {
-
writeJobDetailOverviewAsJson(detail, gen, now);
+ for (JobDetails detail
: result.getRunning()) {
+
jobDetailsSerializer.serialize(detail, gen, null);
}
gen.writeEndArray();
gen.writeArrayFieldStart("finished");
- for (JobDetails detail
: result.getFinishedJobs()) {
-
writeJobDetailOverviewAsJson(detail, gen, now);
+ for (JobDetails detail
: result.getFinished()) {
+
jobDetailsSerializer.serialize(detail, gen, null);
}
gen.writeEndArray();
} else {
--- End diff --
Yes, because I think it's not so nice to return different answer formats
depending on how the handler was instantiated. The respective information can
be easily extracted from the complete overview. Moreover, I think we never used
these endpoints in the first place.
---
