Add 2020 meeting notes archive

This commit is contained in:
Elana Hashman 2021-01-07 10:27:59 -08:00
parent 3f4f86ac1e
commit 2b9c5e9ca3
No known key found for this signature in database
GPG Key ID: D37F7B2A20B48FA0
1 changed files with 679 additions and 0 deletions

View File

@ -0,0 +1,679 @@
## _Triage (2020-12-16)_
Attendees:
* Ehashman
* Logicalhan
* Brancz
* Dashpole
* lilic
* Serathius
* Metalmatze
* Afrouz Mashayekhi
## _Agenda (2020-12-10)_
Issues:
* Last regular SIG meeting of 2020!
* Continue discussion on metrics naming policy
* Dashpole: PR to update the metric naming guideline.
* What’s going on with stable metrics? Lots of bugs against website for the list being empty e.g. [https://github.com/kubernetes/website/issues/24915](https://github.com/kubernetes/website/issues/24915)
* In freeze now, nothing we can do
* SLO support is a concern, reference [https://github.com/kubernetes/community/blob/master/sig-scalability/slos/api_call_latency.md](https://github.com/kubernetes/community/blob/master/sig-scalability/slos/api_call_latency.md)
* Etcd metrics on large numbers of entries is a concern. What is “largeâ€<C3A2> is different per customer
* [lili]: Adding new subproject/or part of kube-state-metrics itself but on another /metrics endpoint, that exposes metrics that do cannot be added to kube-state-metrics, e.g. things that do not map 1:1 to the k8s API [https://github.com/kubernetes/kube-state-metrics/issues/1302#issuecomment-734237720](https://github.com/kubernetes/kube-state-metrics/issues/1302#issuecomment-734237720)
* [lili] kube-state-metrics v2.0.0-beta was cut [https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.0.0-beta](https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.0.0-beta)
Attendees:
*
## _Triage (2020-12-02)_
Issues:
* Test out SIG Node triage tool: [https://docs.google.com/document/d/1JOXKBDgXmQzz8YQSYa7XYcfVteM79iMtvId1aQXC1e8/edit](https://docs.google.com/document/d/1JOXKBDgXmQzz8YQSYa7XYcfVteM79iMtvId1aQXC1e8/edit)
* Kick this to the next triage meeting, probably in the new year
Attendees:
* Ehashman
* Logicalhan
* Joadavis
## _(CANCELLED) Agenda (2020-11-25)_
Cancelled due to US Thanksgiving.
## _(CANCELLED) Triage (2020-11-18)_
Cancelled due to KubeCon.
## _Agenda (2020-11-12)_
Issues:
* [dashpole] FYI Node proxy conformance testing vs removal: [https://github.com/kubernetes/kubernetes/issues/95930](https://github.com/kubernetes/kubernetes/issues/95930). I think prometheus integration depends on this.
* Please add any feedback for groups that may be using these in the issue linked here
* [ehashman] Developer documentation updates [https://github.com/kubernetes/community/issues/5233](https://github.com/kubernetes/community/issues/5233)
* Wrappers for metrics framework - needs to documented [https://github.com/kubernetes/community/issues/5080](https://github.com/kubernetes/community/issues/5080)
* Global metrics registry - Lili
* Tracing - dashpole
* More detailed naming policies for metrics
* Any needed updates for existing docs - logicalhan
* [serathius] Strategy and direction for metrics server [https://github.com/kubernetes-sigs/metrics-server/issues/627](https://github.com/kubernetes-sigs/metrics-server/issues/627)
* Reach out to contribex to run a user survey?
* Discussion about metric naming
* Scheduler: [https://github.com/kubernetes/kubernetes/pull/94866](https://github.com/kubernetes/kubernetes/pull/94866)
* Kubelet: [https://github.com/kubernetes/kubernetes/pull/95839](https://github.com/kubernetes/kubernetes/pull/95839)
* [brancz] Suggestion: there are two types of metrics:
* For k8s resources: node/pod/container…
* For k8s components: apiserver/…
* For resource related metrics, it would be better to use the recommended naming e.g. `kube_pod_…`
* [ehashman] worry that the proposal would confuse the people about where to find the metrics definition.
* kube_* used for KSM only currently
* kube_* doesn’t say where a metric is coming from
* We don’t have any documentation for metrics other than KSM docs
* Operators won’t easily be able to understand these metrics
* [ehashman] Announcements/Reminders:
* Code freeze: TODAY Nov. 12 for 1.20
* Doc freeze: Nov. 30
* Upcoming meetings: Nov/Dec cancellations
Attendees:
* Ehashman
* Logicalhan
* Dashpole
* Brancz
* Serathius
* lilic
* Leszek
* Erain
* Joadavis
* Kakkoyun
* Damien Grisonnet
## _Triage (2020-11-04)_
Attendees:
* Dashpole
* Logicalhan
* Ehashman
* Joadavis
* erain
## _Agenda (2020-10-29)_
No agenda items, cancelling.
## _Triage (2020-10-21)_
Attendees:
* Dashpole
* Kakkoyun
* erain
## _Agenda (2020-10-15)_
Issues:
* [dashpole] Metadata standards: [See this comment](https://github.com/kubernetes/enhancements/pull/1458#issuecomment-705842455)
* We have two competing metadata standards: [OpenTelemetry](https://github.com/open-telemetry/opentelemetry-specification/blob/master/specification/resource/semantic_conventions/k8s.md) and [Kubernetes](https://github.com/kubernetes/community/blob/master/contributors/devel/sig-instrumentation/instrumentation.md).
* Should we change? Ask them to change?
* Why did they choose ‘.’ as the namespace delineator?
* [logicalhan] adding context to metrics interfaces so that we can absorb label values via context, rather than tedious, error-prone argument plumbing
* conclusion: we achieved community consensus on adding context to metrics interfaces in component-base, will be backwards compatible with existing usages so we are free to move forward with this w/o a KEP
* [lili] [https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.0.0-alpha.1](https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.0.0-alpha.1) new pre-release towards 2.0 release
Attendees:
* Brancz
* Dashpole
* Lilic
* Logicalhan
* kakkoyun
* erain
## _Triage (2020-10-07)_
Attendees:
* Ehashman
* Dashpole
* Brancz
* Utkarsh
* erain
## _Agenda (2020-10-01)_
Issues:
* [ehashman] Revisiting KEPs targeted for 1.20 before enhancements freeze
* KEP freeze after this weekend
* 7 KEPs currently open: [https://github.com/kubernetes/enhancements/issues?q=is%3Aopen+is%3Aissue+label%3Asig%2Finstrumentation](https://github.com/kubernetes/enhancements/issues?q=is%3Aopen+is%3Aissue+label%3Asig%2Finstrumentation)
* Will track 5 into 1.20, assuming they are merged as implementable by freeze (Oct. 6)
* [lili] v1 klog compatibility in v1.19
* Was not announced well enough?
* Due for a community update this month - should include that
* [SergeyKanzhelev] Help with some PRs review: [https://docs.google.com/document/d/13TZ2CVXIp9BRUbkyUZMn-GDFRTJL-EC2hnQv0lqh_Vs/edit](https://docs.google.com/document/d/13TZ2CVXIp9BRUbkyUZMn-GDFRTJL-EC2hnQv0lqh_Vs/edit)
* -> triage meeting
* Event deduplication PRs - not clear what’s the rules on what’s goes into the event and what not
* [Added container name in the back-off event log · Issue #93636 · kubernetes/kubernetes](https://github.com/kubernetes/kubernetes/pull/93636)
* [fix duplicate FailedMount events when using multiple volumes · Issue #88369 · kubernetes/kubernetes](https://github.com/kubernetes/kubernetes/pull/88369)
* [dashpole] Demo of [systemd_exporter](https://github.com/dashpole/systemd_exporter)
* Povilas’ version of systemd_exporter: [https://github.com/povilasv/systemd_exporter](https://github.com/povilasv/systemd_exporter)
Attendees:
* ehashman
* joadavis
* Dashpole
* brancz
* Leszek Jakubowski
* Serathius
* SergeyKanzhelev
* Metalmatze
* Steph Rifai
* Damien Grissonet
* Clayton
## _Triage (2020-09-23)_
Attendees:
* ehashman
* dashpole
* erain
## _Agenda (2020-09-17)_
Issues:
* [ehashman] KEPs targeted for 1.20
* Enhancements freeze Oct. 6, code freeze Nov. 12
* [https://github.com/kubernetes/enhancements/issues?q=is%3Aopen+is%3Aissue+label%3Asig%2Finstrumentation](https://github.com/kubernetes/enhancements/issues?q=is%3Aopen+is%3Aissue+label%3Asig%2Finstrumentation)
* Reviewed and updated all open KEPs
* [ehashman] Follow-up on e2e test flakiness/state of logging addons
* Metrics grabber test had a race condition, fixed by dashpole with a retry [https://github.com/kubernetes/kubernetes/issues/93688](https://github.com/kubernetes/kubernetes/issues/93688)
* [https://github.com/kubernetes/kubernetes/issues/92731](https://github.com/kubernetes/kubernetes/issues/92731) might be because of rename of pod; erain@ to (learn to) run the e2e test and confirm.
* [https://github.com/kubernetes/kubernetes/issues/93480](https://github.com/kubernetes/kubernetes/issues/93480) Reassign elasticsearch flake to @brancz
* [lili] Migration of [https://github.com/DirectXMan12/k8s-prometheus-adapter](https://github.com/DirectXMan12/k8s-prometheus-adapter)
* [https://github.com/kubernetes/org/issues/2182](https://github.com/kubernetes/org/issues/2182)
* In progress
* Anyone want to join the OWNERS file? @brancz offers to help with reviews
* [lili] kube-state-metrics 2.0.0-alpha pre release -** breaking changes** - please help testing! [https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.0.0-alpha](https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.0.0-alpha)
Attendees:
* ehashman
* brancz
* dashpole
* lilic
* metalmatze
* Marek
* Sergey Kranzhelev
* erain
* 44past4
## _Triage (2020-09-09)_
Attendees:
* ehashman
* brancz
* dashpole
* kakkoyun
* erain
## _Agenda (2020-09-03)_
Issues:
* [chelseychen] - Overview for [New Event API GA Graduation](https://github.com/kubernetes/enhancements/pull/1662)
* [ehashman] - Follow up on last meeting’s topic (e2e tests)
* Dashpole - looked a little, saw a few isolated instances but still nothing big
* Serathius - will have an update by last week
* [ehashman] Does SIG instrumentation want to use feature tracking similar to SIG Node? (from Lauri Apple)
* [https://docs.google.com/document/d/1-NNfuFMh246_ujQpYBCldLQ9hmqEbukaa79Xe1WGwdQ/edit?ts=5f3da0c9#](https://docs.google.com/document/d/1-NNfuFMh246_ujQpYBCldLQ9hmqEbukaa79Xe1WGwdQ/edit?ts=5f3da0c9#)
* -> defer to triage meeting next week
Attendees:
* ehashman
* dashpole
* joadavis
* Keikhara
* Leszek (makdaam)
* erain
* ...
## _Agenda (2020-08-20)_
Issues:
* [logicalhan] Chairs met last week and have started a weekly triage session on Wednesdays
* On the SIG calendar, anyone can attend
* [ehashman] instrumentation e2e test roundup
* See e.g. [https://github.com/kubernetes/kubernetes/issues/93688](https://github.com/kubernetes/kubernetes/issues/93688) (Metrics grabber)
* ~~dashpole said there’s only one flake because the controller didn’t come up; probably not much to fix~~
* [https://github.com/kubernetes/kubernetes/issues/92731](https://github.com/kubernetes/kubernetes/issues/92731) (Stackdriver logging tests)
* [https://github.com/kubernetes/kubernetes/issues/93480](https://github.com/kubernetes/kubernetes/issues/93480) (Elasticsearch logging tests)
* Need to determine some sort of strategy for unmaintained e2e tests
* Elasticsearch addon does not seem to be in great shape, similar Stackdriver addon was removed
* Marek should have the best context on this
* **Action:** find someone to assist with Stackdriver tests
* ehashman suggests we should just remove or find a maintainer for the logging tests
* Ended early [:46] - short meeting because of KubeCon
Attendees:
* ehashman
* dashpole
* logicalhan
* Serathius
* erain
* ...
## _Agenda (2020-08-06)_
Issues:
* Lauri Apple - triage workflow discussion
* [https://github.com/google/triage-party](https://github.com/google/triage-party)
* [Solly Ross, Yuchen Zhou] - demo a new instrumentation tool (promq)
* Temporary link: [https://github.com/logicalhan/instrumentation-tools](https://github.com/logicalhan/instrumentation-tools)
* Frederic - k8s-prometheus-adapter consensus on adoption
Attendees:
* dashpole
* brancz
* kakkoyun
* lilic
* metalmatze
* ehashman
* logicalhan
* joadavis
* Leszek
## _Agenda (2020-07-09)_
## Issues:
* [Frederic] Adopt [https://github.com/DirectXMan12/k8s-prometheus-adapter](https://github.com/DirectXMan12/k8s-prometheus-adapter) as sig-instrumentation project
* Need to include a clear positioning statement, how it relates to metrics-server
* Makes more sense to be in SIG Instrumentation than to try to push to Prometheus community
[Matthias] Update on [https://github.com/kubernetes/kubernetes/issues/91536](https://github.com/kubernetes/kubernetes/issues/91536)?
Attendees:
* lilic
* brancz
* Leszek
* Kemal Akkoyun
* David Ashpole
* joadavis
* Damien Grisonnet
* Mark Siarkowicz
* Matthias Loibl
## Agenda (2020-06-25)
## Issues:
* [immutableT] [KEP-1753](https://github.com/kubernetes/enhancements/pull/1754): Kubernetes system components logs sanitization
* [serathius] Code freeze moved back by 2 weeks to July 9
Attendees:
* Alextc
* Brancz
* dashpole
* ehashman
* marek
* pawel
* damien grisonnet
* joadavis
* lili
* matthias loibl
* patrick rhomberg
* logicalhan
## Agenda (2020-06-11)
Issues:
* [brancz] [https://github.com/kubernetes/kubernetes/issues/58638](https://github.com/kubernetes/kubernetes/issues/58638)
* [logicalhan] [API call latencies SLO](https://github.com/kubernetes/community/blob/master/sig-scalability/slos/api_call_latency.md) (as we discussed in the previous meeting)
* Should probably also do a sweep through PR backlog
* [immutableT] KEP-1753: Kubernetes system components logs sanitization \
Attendees:
## _ \
_Agenda (2020-05-28)
Attendees:
* logicalhan
* ehashman
* brancz
* dashpole
* rainbowmango
* serathius
* metalmatze
* (this was all i could recollect from memory)
Agenda Issues:
* [logicalhan] setting up SIG repos
* [brancz] Revisit Metric Stability KEP graduation criteria - [https://github.com/kubernetes/kubernetes/issues/91536](https://github.com/kubernetes/kubernetes/issues/91536)
* Kubernetes-mixin already heavily depends on that metric being stable: \
[https://github.com/kubernetes-monitoring/kubernetes-mixin/blob/master/config.libsonnet#L3-L18](https://github.com/kubernetes-monitoring/kubernetes-mixin/blob/master/config.libsonnet#L3-L18) \
[https://github.com/kubernetes-monitoring/kubernetes-mixin/blob/master/alerts/kube_apiserver.libsonnet](https://github.com/kubernetes-monitoring/kubernetes-mixin/blob/master/alerts/kube_apiserver.libsonnet) \
[https://github.com/kubernetes-monitoring/kubernetes-mixin/blob/master/rules/kube_apiserver.libsonnet](https://github.com/kubernetes-monitoring/kubernetes-mixin/blob/master/rules/kube_apiserver.libsonnet)
* [ehashman] (housekeeping) wrap-up of Metrics Overhaul KEP? [https://github.com/kubernetes/enhancements/issues/1206](https://github.com/kubernetes/enhancements/issues/1206)
* Circle back with sig-testing/sig-release to close the loop on this
* **Action:** ehashman to finish KEP
* [logicalhan]
## Agenda (2020-05-14)
Attendees:
* ehashman
* logicalhan
* dashpole
* brancz
* Kakkoyun
* dgrisonnet
* Joadavis
* serathius
Agenda Issues:
* ~~[logicalhan] setting up SIG repos ~~(deferred)
* [44past4] ([New KEP](https://github.com/kubernetes/enhancements/pull/1754)) [Dynamic logs sanitization](https://github.com/kubernetes/enhancements/issues/1753).
* Sensitive data like tokens, keys, passwords can leak in logs for controller-manager, other components
* [logicalhan] [Dynamic cardinality enforcement](https://github.com/kubernetes/enhancements/pull/1692)
* [logicalhan] fixing api request total metric before returning to stable
* Han to file an issue detailing the problems - [https://github.com/kubernetes/kubernetes/issues/91536](https://github.com/kubernetes/kubernetes/issues/91536)
## Agenda (2020-04-30)
Attendees:
* dashpole
* brancz
* ehashman
* logicalhan
* serathius
* liggitt
* joadavis
* lilic
* smarterclayton
* 44past4
* dgrisonnet
Agenda issues:
* [serathius] FYI Enhancement Freeze moved back 2 weeks [kubernetes/sig-release#1065](https://github.com/kubernetes/sig-release/pull/1065)
* [liggitt] [KEP-1693](https://github.com/kubernetes/enhancements/pull/1694): Mechanism for warning API clients about deprecated API use
* proposed deprecated API use metric is relevant to sig-instrumentation
* AI: Jordan follow up with logicalhan/brancz about new metric vs adding to existing request count metric - [https://groups.google.com/g/kubernetes-sig-instrumentation/c/BACdXH4LscY](https://groups.google.com/g/kubernetes-sig-instrumentation/c/BACdXH4LscY)
* [brancz/smarterclayton/lilic] (New KEP) Accurate Pod resource request/limit metric reporting as seen by scheduler. ([Email](https://groups.google.com/forum/?utm_medium=email&utm_source=footer#!topic/kubernetes-sig-instrumentation/Klg_DJxFd54), [Google Doc](https://docs.google.com/document/d/13mjommm_wbwqKtisThmIa33TRa-gfIHJicxQ1S-TOlw/edit#heading=h.scvi1ournpue), [KSM issue](https://github.com/kubernetes/kube-state-metrics/issues/1095))
* AI: smarterclayton to submit a draft KEP for more comment
* [44past4] (New KEP) [Dynamic logs sanitization](https://github.com/kubernetes/enhancements/issues/1753).
* Sensitive data like tokens, keys, passwords can leak in logs for controller-manager, other components
## Agenda (2020-04-16)
Attendees:
* ehashman
* logicalhan
* dashpole
* brancz
* Joadavis
* Serathius
* lilic
Agenda Issues:
* [logicalhan] WIP Cardinality KEP ([https://github.com/kubernetes/enhancements/pull/1692](https://github.com/kubernetes/enhancements/pull/1692))
* Should we plan on bringing stability to GA this release? I would like to start marking metrics as STABLE, especially the ones that we can’t modify without really breaking people.
* Few remaining things from metrics stability KEP: CI checks, dynamic disabling of metrics
* Need to identify metrics proposed to be made STABLE
* Need to contact subproject owners to propose metrics to be made STABLE - someone should draft a communication?
* Is there already some thinking about having a maturity index per metric? Not much other than what is in source control. Metric stability framework should provide some of that, though not GA yet. KEP - [https://github.com/kubernetes/enhancements/blob/master/keps/sig-instrumentation/20190404-kubernetes-control-plane-metrics-stability.md](https://github.com/kubernetes/enhancements/blob/master/keps/sig-instrumentation/20190404-kubernetes-control-plane-metrics-stability.md) Framework will/could have alpha/beta/stable per metric.
* Bug scrub?
* [logicalhan] (related - [https://github.com/kubernetes/kubernetes/issues/89788](https://github.com/kubernetes/kubernetes/issues/89788))
* Fixing the calendar invite [logicalhan]
* Fix meeting notes link
* Invite was created by a coreos account that doesn’t exist any more
## Agenda (2020-04-02)
Attendees:
* Logicalhan
* dashpole
* Akonarde
* kakkoyun
* joadavis
* ehashman
* Lili
* Serathius
* hase1128
* 44past4
Agenda Issues:
* [logicalhan] Cardinality discussion
* Issues:
* [https://github.com/kubernetes/kubernetes/issues/76302](https://github.com/kubernetes/kubernetes/issues/76302)
* [https://github.com/kubernetes/kubernetes/issues/89378](https://github.com/kubernetes/kubernetes/issues/89378)¹
* [https://github.com/kubernetes/kubernetes/issues/89377s
](https://github.com/kubernetes/kubernetes/issues/89377)
* Can/should we introduce any mechanisms to protect against unbounded cardinality?
* Decision: logicalhan will work on a KEP to address this. Should be small
* Bug scrub
## Agenda (2020-03-19)
Attendees:
* logicalhan
* ehashman
* Brancz
* olivierlemasle
* dashpole
* Jpepin
* Metalmatze
* kakkoyun
* Joadavis
* Leszek Jakubowski
* lilic
* Maximillianbrain1
* serathius
Agenda Issues:
* Welcome new tech lead
* David Ashpole
* [ehashman] Fixed terms and/or term limits for SIG leads?
* Proposed: 1 year terms
* ACTION: Han to inquire about setting up lead term limits
* [ehashman] Community update 2020-03-19 done! [Slides](https://docs.google.com/presentation/d/1tWWWsnZZPcoMAYj60jL2541cNA1GyLCR4i-EQzaI9y8/edit#)
* [ehashman] Community meeting -> SIG repos need to be moved out of kubernetes/*
* KSM is affected, may break pretty badly (uses k8s.io imports)
* ACTION: Elana to look for info on repo migration [done]
* Should be an issue in k/org for migration (have done this for metrics-server from k-incubating -> k-sigs)
* Issue: [https://github.com/kubernetes/steering/issues/136](https://github.com/kubernetes/steering/issues/136)
* [metalmatze] I created a PR with an example APIServer SLO implementation (following [SIG Scalability guidelines](https://github.com/kubernetes/community/blob/master/sig-scalability/slos/api_call_latency.md)). Might be interesting for folks and comments welcome (not strictly for this SIG): [https://github.com/kubernetes-monitoring/kubernetes-mixin/pull/382](https://github.com/kubernetes-monitoring/kubernetes-mixin/pull/382)
## Agenda (2020-03-05)
* Welcome new chairs
* logicalhan
* ehashman
* New chairs to set up the vote for Tech Leads
## Agenda (2020-02-20)
* [pszczesniak] Proposal - @serathius to replace @piosz as SIG Chair (see [details](https://groups.google.com/g/kubernetes-sig-instrumentation/c/Rod5jxVluT0))
* Suggestion (ehashman): governance structure change from 2 chairs to up to 2 chairs, up to 2 TLs.
* Suggestion (logicalhan): terms (mandatory re-election on a regular interval) for chairs/TLs.
* Action items:
* Please ensure you are in the [member tracking yml](https://github.com/kubernetes/org/blob/master/config/kubernetes/sig-instrumentation/teams.yaml), and send out email to mailing list
* Call for nominations for new chair/TL positions on mailing list (nominees must confirm their acceptance)
* Secret vote (CIVS?)
## Agenda (2020-02-06)
* [serathius] Update on structured logging
## Agenda (2020-01-23)
* [mark peek] Cloud events introduction
* [hase1128] [I considered how to proceed Request-ID again, so I want feedback.](https://kubernetes.slack.com/archives/C20HH14P7/p1579742164003900)
## Agenda (2020-01-09)
* [dashpole] Discuss how [tracing](https://github.com/kubernetes/enhancements/pull/650), [structured logging](https://github.com/kubernetes/enhancements/pull/1367), and [RequestID](https://github.com/kubernetes/enhancements/pull/1348) proposals interact. \
[Slides with potential solutions](https://docs.google.com/presentation/d/1NmMKaZwIhdgiktSt1OyAktWPdO5D83br0ueVWkDO8uQ/edit?usp=sharing). \
Background: Kubecon tracing [presentation](https://www.youtube.com/watch?v=lEACvRW6T_U)
* [hase1128] [Update Request-ID KEP and add basic interaction idea](https://github.com/hase1128/enhancements/blob/kep-request-id/keps/sig-auth/20191101-add-request-id-to-k8s-logs.md#proposal)
* [brancz] resource metrics API working set defined but metrics-server exposes RSS because VPA [https://github.com/kubernetes-sigs/metrics-server/pull/391](https://github.com/kubernetes-sigs/metrics-server/pull/391)