Parallel document uploads on same bot instance seem to be serialized / slower than expected + how to track upload/download progress?

[mohammadhoseinomidi1387-netizen]

Hi,

I'm running telegram-bot-api (local mode) inside Docker (aiogram/telegram-bot-api:latest image) on a dedicated Hetzner server with good CPU, RAM, and network bandwidth. I have two questions:

1) Concurrent uploads from the same bot seem to be partially serialized

When I send two large files (700MB each) to the same bot concurrently via sendDocument, I observe:

• Single file upload: ~55 seconds
• Two files sent concurrently: first finishes in ~55s, second finishes in ~91s (not ~55s in parallel, but also not ~110s fully sequential)

Test command used:

curl -F "chat_id=<CHAT_ID>" -F "document=@file1.mp4" "http://localhost:8081/bot<TOKEN>/sendDocument" &
curl -F "chat_id=<CHAT_ID>" -F "document=@file2.mp4" "http://localhost:8081/bot<TOKEN>/sendDocument" &
wait

I ruled out network/disk bottlenecks (Hetzner dedicated server, good upload bandwidth, fast local disk). No CPU/memory limits are set on the Docker container. No custom flags are passed (default TELEGRAM_LOCAL=true, TELEGRAM_API_ID, TELEGRAM_API_HASH only).

Is this expected behavior due to how TDLib handles the internal thread pool / actor model per bot instance? Is there a way to increase parallelism for uploads from a single bot (e.g. via --cpu-affinity, --max-connections, or another flag), or is queuing per-bot instance expected by design?

2) How can I track upload/download progress of a file in real time?

Is there any API method, webhook event, or getFile-style polling mechanism that reports the current bytes uploaded/downloaded for an in-progress file transfer (similar to TDLib's updateFile with local/remote FileDownload/UploadedSize)? I'd like to show progress percentage to end users while a large file is being uploaded or downloaded through the Bot API server.

Thanks in advance!