@tvcantho: Săn ‘thủy quái’ rừng tràm ở U Minh Thượng

Cần Thơ - Báo & PTTH

Open In TikTok:

Region: VN

Saturday 09 May 2026 13:50:58 GMT

17469

221

7

16

Music

Download

No Watermark .mp4 (14.11MB) No Watermark(HD) .mp4 (21.87MB) Watermark .mp4 (16.33MB) Music .mp3

Comments

Duy 68 :

ở bìa rừng đây ạ 🥰

2026-05-09 14:05:31

1

Thơ Nguyễn 68 Kiên Giang :

ca nhiều lắm

2026-05-09 14:07:10

0

phạm oanh :

cho hỏi tâm ngàn hay 8 chục ngàn

2026-05-09 13:58:46

0

Ahihi :

2026-05-09 13:56:51

0

Thơ Nguyễn 68 Kiên Giang :

quê em đó

2026-05-09 14:06:50

0

Ueirt :

cá lóc = thủy quái 🤣🤣🤣

2026-05-09 13:59:53

0

thanhlam :

😂😂😂

2026-05-10 12:27:43

0

To see more videos from user @tvcantho, please go to the Tikwm homepage.

Other Videos

Chịu chơi, chịu chi quá rồi gleeglee ơi #a #canhgabachthaogleeglee #a #changarutxuonggleeglee #changarutxuong #changadahovisate #changadaho #gleeglee

Chịu chơi, chịu chi quá rồi gleeglee ơi #a #canhgabachthaogleeglee #a #changarutxuonggleeglee #changarutxuong #changadahovisate #changadaho #gleeglee

#fyp

Your Airflow DAG failed at 2:00 AM. Does an engineer need to wake up first? 🚨 This is Graph Engineering in production: Airflow alert → incident router → specialist agents → shared evidence → safe retry → verified recovery ✅ Instead of one AI agent guessing everything, a crew investigates in parallel: 🟢 Log Agent checks errors and stack traces 🔵 Data Agent validates upstream dependencies 🟡 Infra Agent checks memory, CPU and infrastructure 🟣 Change Agent reviews recent deployments Their findings are stored in a shared incident state containing evidence, hypotheses and retry history. Then comes the critical decision: Is the failure transient—and is the task safe to retry? ✅ Yes → rerun only the failed task 🛑 No → pause and escalate to the human on-call But a successful retry is not enough. The system verifies that the task passed, downstream dependencies are healthy and the DAG is green before closing the incident. That is the difference between blind automation and trustworthy agentic operations: ❌ Retry everything ❌ Let one agent improvise ❌ Close after a command succeeds ✅ Route by failure type ✅ Preserve shared incident state ✅ Use bounded retries ✅ Escalate uncertainty ✅ Close only after verification The future of on-call isn’t “AI replaces engineers.” It’s AI handling repetitive diagnosis and safe recovery while humans control ambiguous, high-risk decisions. 🧠 Would you trust this system with your next Airflow failure? 👇 Follow @hackproduct for practical AI engineering, data engineering and system design. #Airflow #DataEngineering #AIEngineering #AIAgents #AgenticAI

Your Airflow DAG failed at 2:00 AM. Does an engineer need to wake up first? 🚨 This is Graph Engineering in production: Airflow alert → incident router → specialist agents → shared evidence → safe retry → verified recovery ✅ Instead of one AI agent guessing everything, a crew investigates in parallel: 🟢 Log Agent checks errors and stack traces 🔵 Data Agent validates upstream dependencies 🟡 Infra Agent checks memory, CPU and infrastructure 🟣 Change Agent reviews recent deployments Their findings are stored in a shared incident state containing evidence, hypotheses and retry history. Then comes the critical decision: Is the failure transient—and is the task safe to retry? ✅ Yes → rerun only the failed task 🛑 No → pause and escalate to the human on-call But a successful retry is not enough. The system verifies that the task passed, downstream dependencies are healthy and the DAG is green before closing the incident. That is the difference between blind automation and trustworthy agentic operations: ❌ Retry everything ❌ Let one agent improvise ❌ Close after a command succeeds ✅ Route by failure type ✅ Preserve shared incident state ✅ Use bounded retries ✅ Escalate uncertainty ✅ Close only after verification The future of on-call isn’t “AI replaces engineers.” It’s AI handling repetitive diagnosis and safe recovery while humans control ambiguous, high-risk decisions. 🧠 Would you trust this system with your next Airflow failure? 👇 Follow @hackproduct for practical AI engineering, data engineering and system design. #Airflow #DataEngineering #AIEngineering #AIAgents #AgenticAI

🧏🏻‍♀️ #capcut#xh

🧏🏻‍♀️ #capcut#xh

The kids were scared… then it got even worse 😭😂 Be honest. Are you getting in or waiting for the next elevator? 👀 #dog #fyp #canecorso

The kids were scared… then it got even worse 😭😂 Be honest. Are you getting in or waiting for the next elevator? 👀 #dog #fyp #canecorso

About

Robot
API

Legal

Privacy Policy