This is an automated email from the ASF dual-hosted git repository.
chia7712 pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/yunikorn-site.git
The following commit(s) were added to refs/heads/master by this push:
new fffcac7ceb [YUNIKORN-2443] Update the features page with the event
system (#404)
fffcac7ceb is described below
commit fffcac7ceb5d68627fc5d616932a45a44482f7f7
Author: Peter Bacsko <[email protected]>
AuthorDate: Thu Mar 21 11:18:38 2024 +0800
[YUNIKORN-2443] Update the features page with the event system (#404)
Closes: #404
Signed-off-by: Chia-Ping Tsai <[email protected]>, Craig Condit
<[email protected]>
---
docs/api/scheduler.md | 2 +-
docs/get_started/core_features.md | 6 ++++++
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/docs/api/scheduler.md b/docs/api/scheduler.md
index 677a64a469..2c0f2c98c4 100644
--- a/docs/api/scheduler.md
+++ b/docs/api/scheduler.md
@@ -2105,7 +2105,7 @@ Endpoint is used to retrieve a batch of event records.
**Code** : `500 Internal Server Error`
-### Event stream
+## Event stream
Creates a persistent HTTP connection for event streaming. New events are sent
to the clients immediately, so unlike the batch interface, there is no need for
polling.
The number of active connections is limited. The default setting is 100
connections total and 15 connections per host. The respective configmap
properties are `event.maxStreams` and `event.maxStreamsPerHost`.
diff --git a/docs/get_started/core_features.md
b/docs/get_started/core_features.md
index ca9f701201..e865e68f91 100644
--- a/docs/get_started/core_features.md
+++ b/docs/get_started/core_features.md
@@ -93,3 +93,9 @@ The [Partition and Queue
Configuration](user_guide/queue_config.md) provides con
## CPU Architecture support
YuniKorn supports running on ARM as well as on AMD/Intel CPUs.
With the release of YuniKorn 1.1.0, prebuilt convenience images for both
architectures are provided in docker hub.
+
+## Event system and application history tracking
+Whenever something relevant happens inside the scheduler (eg. request is
added, allocation happens in a queue, node removal, queue or user quota is
exceeded, etc) an appropriate event is generated.
+This enables users to reliably generate usage statistics about their workloads
over time. Since the event structure is well-defined, this is much more
suitable for automated processing than logs.
+Application history (state transitions & allocations), node, queue, user and
group resource usage can be examined without having to look at the logs.
+The events accessible on the REST interface in two ways: batch or streaming. A
[batch query](../api/scheduler.md#batch-events) simply returns the list of
events, while the [streaming](../api/scheduler.md#event-stream) endpoint keeps
the connection open and new events are immediately sent to the client.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
