- Fixed a race condition in the CEL compiler that could occur when initializing composited policies concurrently. 
  
  ### Description
  Fixes a fatal crash (concurrent map read/write) in `NewCompositedCompilerFromTemplate`.
  

# optionally, remove the old secret
kubectl delete secret --namespace=federation federation-apiserver-secret
```

- Kubernetes components no longer handle panics, and instead actively crash.  All Kubernetes components should be run by something that actively restarts them. This is true of the default setups, but those with custom environments may need to double-check (#28800, @lavalamp)

- Kube-scheduler: fixes a 1.27.14 regression that can lead to a scheduler crash when processing pods with affinity that doesn't match a real/valid node ([#125043](https://github.com/kubernetes/kubernetes/pull/125043), [@AxeZhan](https://github.com/AxeZhan)) [SIG Scheduling and Testing]

## Dependencies

- Fixed potential scheduler crash when scheduling with unsatisfied nodes in PodTopologySpread. ([#110984](https://github.com/kubernetes/kubernetes/pull/110984), [@kerthcet](https://github.com/kerthcet)) [SIG Scheduling]

- Kube-apiserver: fixes a potential crash serving CustomResourceDefinitions that combine an invalid schema and CEL validation rules. ([#126167](https://github.com/kubernetes/kubernetes/pull/126167), [@cici37](https://github.com/cici37)) [SIG API Machinery and Testing]

    assertThat(get.headers["User-Agent"]).isEqualTo("App 1.0")
    assertThat(hostnameVerifier.calls).containsExactly("verify android.com")
  }

  /**
   * We had a bug where OkHttp would crash if HTTP proxies returned a truncated response.
   * https://github.com/square/okhttp/issues/5727
   */
  @Test
  fun proxyUpgradeFailsWithTruncatedResponse() {
    server.enqueue(
      MockResponse

- Kube-apiserver: fixes a potential crash serving CustomResourceDefinitions that combine an invalid schema and CEL validation rules. ([#126167](https://github.com/kubernetes/kubernetes/pull/126167), [@cici37](https://github.com/cici37)) [SIG API Machinery and Testing]

    *   reflector_watches_total
    * While this is a backwards-incompatible change, it would have been impossible to setup reliable monitoring around these metrics since the labels were not stable. 
* Fix keymutex issues which may crash in some platforms. ([#74386](https://github.com/kubernetes/kubernetes/pull/74386), [@danielqsj](https://github.com/danielqsj))



# v1.12.6

[Documentation](https://docs.k8s.io)

## Downloads for v1.12.6

- Fix memory leak on kube-scheduler preemption ([#111803](https://github.com/kubernetes/kubernetes/pull/111803), [@amewayne](https://github.com/amewayne)) [SIG Scheduling]
- Fixed potential scheduler crash when scheduling with unsatisfied nodes in PodTopologySpread. ([#111511](https://github.com/kubernetes/kubernetes/pull/111511), [@kerthcet](https://github.com/kerthcet)) [SIG Scheduling]

しかしながら、**アプリケーション全体をクラッシュさせるようなコードを書いて**UvicornとPythonをクラッシュさせるようなケースもあるかもしれません。💥

それでも、ある箇所でエラーが発生したからといって、アプリケーションを停止させたままにしたくないでしょう。 少なくとも壊れていない*path operation*については、**実行し続けたい**はずです。

### クラッシュ後の再起動 { #restart-after-crash }

しかし、実行中の**プロセス**をクラッシュさせるような本当にひどいエラーの場合、少なくとも2〜3回ほどプロセスを**再起動**させる外部コンポーネントが必要でしょう。

/// tip | 豆知識

...とはいえ、アプリケーション全体が**すぐにクラッシュする**のであれば、いつまでも再起動し続けるのは意味がないでしょう。しかし、その場合はおそらく開発中か少なくともデプロイ直後に気づくと思われます。