BUG: "recordId" value is ignored by batch inference job

The batch inference job results started to use simple incremented numbers instead of provided recordId values, some time between May 7 and May 9.

To reproduce the issue

1. I uploaded the job inference data to s3. First 4 lines preview:

{"recordId":"60a278b24ce","modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}}}
{"recordId":"8d0b82473ee","modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}}}
{"recordId":"f3e12e9f4af","modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}}}
{"recordId":"91194989b01","modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}}}

2. I started the batch inference job with eu.amazon.nova-micro-v1:0
3. I waited when the corresponding job switched into COMPLETED status
4. I downloaded the results

{"modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}},"modelOutput":{"output":{"message":{"content":[{"text":"Sanitised"}],"role":"assistant"}},"stopReason":"end_turn","usage":{"inputTokens":731,"outputTokens":69,"totalTokens":800,"cacheReadInputTokenCount":0,"cacheWriteInputTokenCount":0}},"recordId":"1"}
{"modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}},"modelOutput":{"output":{"message":{"content":[{"text":"Sanitised"}],"role":"assistant"}},"stopReason":"end_turn","usage":{"inputTokens":731,"outputTokens":69,"totalTokens":800,"cacheReadInputTokenCount":0,"cacheWriteInputTokenCount":0}},"recordId":"2"}
{"modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}},"modelOutput":{"output":{"message":{"content":[{"text":"Sanitised"}],"role":"assistant"}},"stopReason":"end_turn","usage":{"inputTokens":731,"outputTokens":69,"totalTokens":800,"cacheReadInputTokenCount":0,"cacheWriteInputTokenCount":0}},"recordId":"3"}
{"modelInput":{"schemaVersion":"messages-v1","messages":[{"role":"user","content":[{"text":"Sanitised"}]}],"system":[{"text":"Sanitised"}],"inferenceConfig":{"maxTokens":256,"topK":1,"stopSequences":["}"]}},"modelOutput":{"output":{"message":{"content":[{"text":"Sanitised"}],"role":"assistant"}},"stopReason":"end_turn","usage":{"inputTokens":731,"outputTokens":69,"totalTokens":800,"cacheReadInputTokenCount":0,"cacheWriteInputTokenCount":0}},"recordId":"4"}

Notice that the "recordId" values are just strings with natural numbers starting from "1"

I've noticed that some of the documentation pages refer to the recordId just as "alphanumeric string" while others mention "11 character alphanumeric string". Even though strings of any length worked previously, I decided to use 11 char string for this specific experiment.

I've also noticed that similarly initiated job created on May 7 did not suffer from the same issue, so I suppose there were some updates between these dates.