Add event exporter deployment to the fluentd-gcp addon #46700
Conversation
crassirostris
added
the
sig/instrumentation
k8s-ci-robot
added
the
cncf-cla: yes
k8s-github-robot
added
approved
| command: | ||
| - '/event-exporter' | ||
| resources: | ||
| limits: |
There was a problem hiding this comment.
Don't we want to autoscale this?
cc @gmarek
There was a problem hiding this comment.
I don't know - you tell me how many events per second it can handle with this settings.
There was a problem hiding this comment.
Yes, we do, I was just about to add it
I was checking the throughput to decide whether cpu has to be scaled. Manual test with 100 events/sec showed that cpu is not a bottleneck, but memory is
@gmarek Is it safe to assume that 50MB + 1MB/node covers the size of the events db in typical cluster?
There was a problem hiding this comment.
You mean db as in etcd? For 500 Nodes it's <200MB, so this is very safe estimate. You probably can do .5MB/node. But that's etcd - I'm not sure how this related to your needs though.
There was a problem hiding this comment.
@gmarek Great, thanks. I've observed that the db size when event exporter crashed was ~ the same size as the limit
|
@roberthbailey could you please approve or delegate? |
crassirostris
force-pushed
the
add-event-exporter-deployment
branch
2 times, most recently
from
8c68fd9 to
f95024a
Compare
k8s-github-robot
added
size/L
|
@piosz PTAL at the nanny config |
crassirostris
force-pushed
the
add-event-exporter-deployment
branch
from
f95024a to
07902ea
Compare
|
Test failures are flakes due to missing quota in the test project: #46713 |
|
@k8s-bot pull-kubernetes-kubemark-e2e-gce test this |
| kubernetes.io/cluster-service: "true" | ||
| addonmanager.kubernetes.io/mode: Reconcile | ||
| --- | ||
| apiVersion: rbac.authorization.k8s.io/v1alpha1 |
There was a problem hiding this comment.
Please use rbac.authorization.k8s.io/v1beta1 instead.
There was a problem hiding this comment.
Thanks for noticing! Done
crassirostris
force-pushed
the
add-event-exporter-deployment
branch
from
07902ea to
238db08
Compare
|
/approve |
| fieldRef: | ||
| fieldPath: metadata.namespace | ||
| command: | ||
| - /pod_nanny |
There was a problem hiding this comment.
The number of pod_nannies is getting larger and larger.
Can we file an issue to make it possible to use a single pod nanny for many different pods?
crassirostris
force-pushed
the
add-event-exporter-deployment
branch
2 times, most recently
from
b82a482 to
d88ea16
Compare
k8s-github-robot
added
size/M
|
@k8s-bot pull-kubernetes-federation-e2e-gce test this |
|
@k8s-bot pull-kubernetes-e2e-kops-aws test this |
|
@piosz I removed resources to make them BestEffort |
crassirostris
force-pushed
the
add-event-exporter-deployment
branch
from
d88ea16 to
88d9e55
Compare
crassirostris
force-pushed
the
add-event-exporter-deployment
branch
from
88d9e55 to
527206c
Compare
|
@k8s-bot pull-kubernetes-e2e-kops-aws test this |
|
/lgtm |
k8s-ci-robot
added
the
lgtm
|
@k8s-bot pull-kubernetes-e2e-kops-aws test this |
|
@k8s-bot pull-kubernetes-verify test this |
|
@k8s-bot pull-kubernetes-unit test this |
|
@k8s-bot pull-kubernetes-federation-e2e-gce test this |
k8s-github-robot
removed
the
approved
|
/approve no-issue |
2 similar comments
|
/approve no-issue |
|
/approve no-issue |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: crassirostris, piosz, roberthbailey Associated issue requirement bypassed by: piosz The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these OWNERS Files:
You can indicate your approval by writing |
k8s-github-robot
added
the
approved
|
@crassirostris: The following test failed, say
Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
Automatic merge from submit-queue |
Introduce event exporter deployment to the fluentd-gcp addon so that by default if logging to Stackdriver is enabled, events will be available there also.
In this release, event exporter is a non-critical pod in BestEffort QoS class to avoid preempting actual workload in tightly loaded clusters. It will become critical in one of the future releases.