gnodet-bot commented on PR #27352:
URL: https://github.com/apache/camel/pull/27352#issuecomment-5994178740
## ❌ Local IT results for `e505a459d8a6045426cdc3e8ceb41b70ea0ea48d`
Integration tests run locally (without `-Dci.env.name`) against the upgraded
`docling.container v1.36.0` image revealed **performance regressions** causing
timeout failures.
**Test-infra module**: `camel-test-infra-docling`
**Component tested**: `components/camel-ai/camel-docling`
### Results
| Test Class | Tests | Failures | Notes |
|---|---|---|---|
| `ExtractStructuredDataIT` | 3 | ❌ 1 | `extractStructuredDataFromInvoice`
timed out after 122s |
| `OcrExtractionIT` | 6 | ❌ 1 | `testOcrWithMultipleTextBlocks` timed out
after 120s |
| `ChunkingIT` | 6 | ✅ 0 | |
| `BatchProcessingIT` | — | ⚠️ N/A | Killed by Maven shutdown hook (cascade
from fork timeout) |
| `DoclingServeProducerIT` | — | ⚠️ N/A | Not reached |
| `MetadataExtractionIT` | — | ⚠️ N/A | Not reached |
### Root Cause
Both failing tests hit the 120s `processTimeout` configured in
`DoclingITestSupport`:
- `extractStructuredDataFromInvoice`: complex PDF invoice → sync extraction
times out at 120s
- `testOcrWithMultipleTextBlocks`: OCR on multi-block image → also times out
at 120s
These same tests passed with `v1.35.0`. The new `v1.36.0` image appears to
be slower for heavy processing tasks (invoice PDF analysis, multi-block OCR),
exceeding the 120 000 ms client-side request timeout.
This is a **non-trivial behavioral regression** that requires investigation:
1. Is `v1.36.0` genuinely slower, or does it require more warmup?
2. Should `processTimeout` in the ITs be increased (e.g. to 300s)?
3. Should the test be changed to use async mode with longer poll timeout?
Not merging until this is understood. Flagging for human review.
<details><summary>Failing test output (ExtractStructuredDataIT)</summary>
```
[ERROR] Tests run: 3, Failures: 0, Errors: 1, Skipped: 0, Time elapsed:
210.9 s <<< FAILURE!
[ERROR]
org.apache.camel.component.docling.integration.ExtractStructuredDataIT.extractStructuredDataFromInvoice
-- Time elapsed: 122.1 s <<< ERROR!
Caused by: ai.docling.serve.client.DoclingServeClientException: request
timed out
at
ai.docling.serve.client.DoclingServeClient.execute(DoclingServeClient.java:232)
...
Caused by: java.net.http.HttpTimeoutException: request timed out
```
</details>
<details><summary>Failing test output (OcrExtractionIT)</summary>
```
[ERROR] Tests run: 6, Failures: 0, Errors: 1, Skipped: 0, Time elapsed:
318.3 s <<< FAILURE!
[ERROR]
org.apache.camel.component.docling.integration.OcrExtractionIT.testOcrWithMultipleTextBlocks
-- Time elapsed: 120.2 s <<< ERROR!
Caused by: java.io.IOException: Failed to convert document: request timed out
...
Caused by: ai.docling.serve.client.DoclingServeClientException: request
timed out
```
</details>
> _Local integration tests run by Hermes on behalf of @gnodet_
> _Tests run without `-Dci.env.name` to include ITs disabled on GitHub
Actions CI_
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
