|
24 | 24 | - [Graduation Criteria](#graduation-criteria)
|
25 | 25 | - [Upgrade / Downgrade Strategy](#upgrade--downgrade-strategy)
|
26 | 26 | - [Version Skew Strategy](#version-skew-strategy)
|
| 27 | +- [Production Readiness Review Questionnaire](#production-readiness-review-questionnaire) |
| 28 | + - [Feature Enablement and Rollback](#feature-enablement-and-rollback) |
| 29 | + - [Rollout, Upgrade and Rollback Planning](#rollout-upgrade-and-rollback-planning) |
| 30 | + - [Monitoring Requirements](#monitoring-requirements) |
| 31 | + - [Dependencies](#dependencies) |
| 32 | + - [Scalability](#scalability) |
27 | 33 | - [Implementation History](#implementation-history)
|
28 | 34 | <!-- /toc -->
|
29 | 35 |
|
30 | 36 | ## Release Signoff Checklist
|
31 | 37 |
|
32 |
| -- [x] kubernetes/enhancements issue in release milestone, which links to KEP (this should be a link to the KEP location in kubernetes/enhancements, not the initial KEP PR) |
33 |
| - - [#859](https://github.com/kubernetes/enhancements/issues/859) |
34 |
| -- [x] KEP approvers have set the KEP status to `implementable` |
35 |
| -- [x] Design details are appropriately documented |
36 |
| -- [x] Test plan is in place, giving consideration to SIG Architecture and SIG Testing input |
37 |
| - - Standard Unit and Integration testing should be sufficient |
38 |
| -- [x] Graduation criteria is in place |
39 |
| - - This is not a user facing API change |
| 38 | +Items marked with (R) are required *prior to targeting to a milestone / release*. |
| 39 | + |
| 40 | +- [X] (R) Enhancement issue in release milestone, which links to KEP dir in [kubernetes/enhancements] (not the initial KEP PR) |
| 41 | +- [X] (R) KEP approvers have approved the KEP status as `implementable` |
| 42 | +- [X] (R) Design details are appropriately documented |
| 43 | +- [X] (R) Test plan is in place, giving consideration to SIG Architecture and SIG Testing input |
| 44 | +- [X] (R) Graduation criteria is in place |
| 45 | +- [X] (R) Production readiness review completed |
| 46 | +- [X] (R) Production readiness review approved |
40 | 47 | - [ ] "Implementation History" section is up-to-date for milestone
|
41 | 48 | - [ ] User-facing documentation has been created in [kubernetes/website], for publication to [kubernetes.io]
|
42 |
| -- [ ] Supporting documentation e.g., additional design documents, links to mailing list discussions/SIG meetings, relevant PRs/issues, release notes |
| 49 | +- [ ] Supporting documentation—e.g., additional design documents, links to mailing list discussions/SIG meetings, relevant PRs/issues, release notes |
| 50 | + |
43 | 51 |
|
44 | 52 | [kubernetes.io]: https://kubernetes.io/
|
45 | 53 | [kubernetes/enhancements]: https://github.com/kubernetes/enhancements/issues
|
@@ -257,11 +265,175 @@ included.
|
257 | 265 |
|
258 | 266 | ### Upgrade / Downgrade Strategy
|
259 | 267 |
|
260 |
| -NA |
| 268 | +Not applicable. There are no cluster components affected by this feature. |
| 269 | + |
261 | 270 |
|
262 | 271 | ### Version Skew Strategy
|
263 | 272 |
|
264 |
| -NA |
| 273 | +Not applicable. There is nothing required of the API Server, so there |
| 274 | +can be no version skew. |
| 275 | + |
| 276 | +## Production Readiness Review Questionnaire |
| 277 | + |
| 278 | +### Feature Enablement and Rollback |
| 279 | + |
| 280 | +* **How can this feature be enabled / disabled in a live cluster?** |
| 281 | + |
| 282 | + This feature is enabled with the KUBECTL_COMMANDS_HEADER environment |
| 283 | + variable set on the client command line. Since it only affects the client |
| 284 | + kubectl, it does not affect any cluster components. |
| 285 | + |
| 286 | + - [X] Feature gate (also fill in values in `kep.yaml`) |
| 287 | + - Feature gate name: KUBECTL_COMMANDS_HEADERS |
| 288 | + - Components depending on the feature gate: kubectl |
| 289 | + - [X] Other |
| 290 | + - Describe the mechanism: setting an client-side environment variable for kubectl. |
| 291 | + - Will enabling / disabling the feature require downtime of the control |
| 292 | + plane? No |
| 293 | + - Will enabling / disabling the feature require downtime or reprovisioning |
| 294 | + of a node? (Do not assume `Dynamic Kubelet Config` feature is enabled). |
| 295 | + No. |
| 296 | + |
| 297 | +* **Does enabling the feature change any default behavior?** |
| 298 | + |
| 299 | + No. This feature is not user facing, so it does not change behavior. |
| 300 | + |
| 301 | +* **Can the feature be disabled once it has been enabled (i.e. can we roll back |
| 302 | + the enablement)?** |
| 303 | + |
| 304 | + Yes. This feature can be disabled by simply removing an environment variable |
| 305 | + on the client command line. |
| 306 | + |
| 307 | +* **What happens if we reenable the feature if it was previously rolled back?** |
| 308 | + |
| 309 | + Re-enabling this feature is simply accomplished by setting the feature |
| 310 | + environment variable on the client command line. There is no state, and there |
| 311 | + is no consequence for re-enabling the feature. |
| 312 | + |
| 313 | +* **Are there any tests for feature enablement/disablement?** |
| 314 | + |
| 315 | + There will be unit tests and integration tests which test this |
| 316 | + feature enablement and disablement by setting the environment |
| 317 | + variable. |
| 318 | + |
| 319 | +### Rollout, Upgrade and Rollback Planning |
| 320 | + |
| 321 | +* **How can a rollout fail? Can it impact already running workloads?** |
| 322 | + |
| 323 | + A danger in the rollout of this feature is adding too much data (too many |
| 324 | + headers) to each of the API Server calls. We intend to mitigate this risk |
| 325 | + by defining a MAX_HEADERS concept to ensure the headers to not grow above a |
| 326 | + certain size. MAX_HEADERS will provide a mechanism to ensure the headers data |
| 327 | + does not grow without bound. As far as cluster workloads, this feature only |
| 328 | + affects kubectl; not any cluster components. So it would not be possible to |
| 329 | + impact running workloads. |
| 330 | + |
| 331 | +* **What specific metrics should inform a rollback?** |
| 332 | + |
| 333 | + We will measure the **headers-added-round-trip-time** and compare it to the round-trip |
| 334 | + time for the same API Server call without headers. This ratio will give us the |
| 335 | + performance penalty for adding these headers. If this performance penalty exceedes |
| 336 | + a specific threshold, users can opt-out by removing the client-side command line |
| 337 | + feature environment variable to disable the headers. |
| 338 | + |
| 339 | +* **Were upgrade and rollback tested? Was the upgrade->downgrade->upgrade path tested?** |
| 340 | + |
| 341 | + Enabling and disabling the client-side command line environment variable will |
| 342 | + be tested. |
| 343 | + |
| 344 | +* **Is the rollout accompanied by any deprecations and/or removals of features, APIs, |
| 345 | +fields of API types, flags, etc.?** |
| 346 | + |
| 347 | + No |
| 348 | + |
| 349 | +### Monitoring Requirements |
| 350 | + |
| 351 | +_This section must be completed when targeting beta graduation to a release._ |
| 352 | + |
| 353 | +* **How can an operator determine if the feature is in use by workloads?** |
| 354 | + Ideally, this should be a metric. Operations against the Kubernetes API (e.g., |
| 355 | + checking if there are objects with field X set) may be a last resort. Avoid |
| 356 | + logs or events for this purpose. |
| 357 | + |
| 358 | + Operators will see this feature in use when the **X-Kubectl-Command** header |
| 359 | + arrives with REST calls to the API Server. |
| 360 | + |
| 361 | +* **What are the SLIs (Service Level Indicators) an operator can use to determine |
| 362 | +the health of the service?** |
| 363 | + |
| 364 | + Since the cardinality is low, we will initially support the kubectl subcommand |
| 365 | + header (X-Kubectl-Command) in this feature. This simple metric will provide valuable |
| 366 | + insight into kubectl usage, and allow users to see if the feature is being used. |
| 367 | + |
| 368 | + - [X] Metrics |
| 369 | + - Metric name: X-Kubectl-Command header |
| 370 | + - [Optional] Aggregation method: |
| 371 | + - Components exposing the metric: kubectl |
| 372 | + - [ ] Other (treat as last resort) |
| 373 | + - Details: |
| 374 | + |
| 375 | +* **What are the reasonable SLOs (Service Level Objectives) for the above SLIs?** |
| 376 | + |
| 377 | + TBD |
| 378 | + |
| 379 | +* **Are there any missing metrics that would be useful to have to improve observability |
| 380 | +of this feature?** |
| 381 | + |
| 382 | + We believe all the other specified **X-Header** will be useful, but we are starting |
| 383 | + out simply in alpha to prove the value and feasibility of the feature. We will add |
| 384 | + others as we approach beta. |
| 385 | + |
| 386 | +### Dependencies |
| 387 | + |
| 388 | +_This section must be completed when targeting beta graduation to a release._ |
| 389 | + |
| 390 | +* **Does this feature depend on any specific services running in the cluster?** |
| 391 | + Think about both cluster-level services (e.g. metrics-server) as well |
| 392 | + as node-level agents (e.g. specific version of CRI). Focus on external or |
| 393 | + optional services that are needed. For example, if this feature depends on |
| 394 | + a cloud provider API, or upon an external software-defined storage or network |
| 395 | + control plane. |
| 396 | + |
| 397 | + For each of these, fill in the following—thinking about running existing user workloads |
| 398 | + and creating new ones, as well as about cluster-level services (e.g. DNS): |
| 399 | + - [Dependency name] |
| 400 | + - Usage description: |
| 401 | + - Impact of its outage on the feature: |
| 402 | + - Impact of its degraded performance or high-error rates on the feature: |
| 403 | + |
| 404 | +### Scalability |
| 405 | + |
| 406 | +* **Will enabling / using this feature result in any new API calls?** |
| 407 | + |
| 408 | + No |
| 409 | + |
| 410 | +* **Will enabling / using this feature result in introducing new API types?** |
| 411 | + |
| 412 | + No |
| 413 | + |
| 414 | +* **Will enabling / using this feature result in any new calls to the cloud |
| 415 | +provider?** |
| 416 | + |
| 417 | + No |
| 418 | + |
| 419 | +* **Will enabling / using this feature result in increasing size or count of |
| 420 | +the existing API objects?** |
| 421 | + |
| 422 | + No |
| 423 | + |
| 424 | +* **Will enabling / using this feature result in increasing time taken by any |
| 425 | +operations covered by [existing SLIs/SLOs]?** |
| 426 | + |
| 427 | + Possibly. This feature increases the size of the REST call from kubectl |
| 428 | + to the API Server by adding more headers to the calls. We will monitor |
| 429 | + this request size increase to ensure there is no deleterious effect. |
| 430 | + |
| 431 | +* **Will enabling / using this feature result in non-negligible increase of |
| 432 | +resource usage (CPU, RAM, disk, IO, ...) in any components?** |
| 433 | + |
| 434 | + Possibly. This feature increases the size of the REST call from kubectl |
| 435 | + to the API Server by adding more headers to the calls. We will monitor |
| 436 | + this request size increase to ensure there is no deleterious effect. |
265 | 437 |
|
266 | 438 | ## Implementation History
|
267 | 439 |
|
|
0 commit comments