ResourceNotFoundException if shard does not exist

[jhecking]

We persist the stream's shard state periodically. When the service restarts, it loads the last known shard state. Occasionally this throws an error if the saved state contains a shard, which no longer exists:

22021-07-02T02:15:08.675Z (node:7) UnhandledPromiseRejectionWarning: ResourceNotFoundException: Requested resource not found: Shard does not exist
32021-07-02T02:15:08.675Z     at deserializeAws_json1_0GetShardIteratorCommandError (/var/app/node_modules/@aws-sdk/client-dynamodb-streams/dist/cjs/protocols/Aws_json1_0.js:240:41)
42021-07-02T02:15:08.675Z     at runMicrotasks (<anonymous>)
52021-07-02T02:15:08.675Z     at processTicksAndRejections (internal/process/task_queues.js:95:5)
62021-07-02T02:15:08.675Z     at async /var/app/node_modules/@aws-sdk/middleware-serde/dist/cjs/deserializerMiddleware.js:6:20
72021-07-02T02:15:08.675Z     at async /var/app/node_modules/@aws-sdk/middleware-signing/dist/cjs/middleware.js:12:24
82021-07-02T02:15:08.675Z     at async StandardRetryStrategy.retry (/var/app/node_modules/@aws-sdk/middleware-retry/dist/cjs/StandardRetryStrategy.js:51:46)
92021-07-02T02:15:08.675Z     at async /var/app/node_modules/@aws-sdk/middleware-logger/dist/cjs/loggerMiddleware.js:6:22
102021-07-02T02:15:08.675Z     at async DynamoDBStream._getShardIterator (/var/app/node_modules/dynamodb-stream/index.js:150:29)
112021-07-02T02:15:08.675Z     at async dispatch (/var/app/node_modules/@kessler/async-map-limit/impl.js:30:22)

I think the client should just ignore such errors and remove the shard from its list.