fix the bug of thread block every request

[MarsYoung]

get_active_containers is blocking every request, make a single request a long response time.

[MarsYoung]

@fengtality @cardosofede @tomasgaudino @rapcmia
Could you please review this PR? It fixes a blocking issue in get_active_containers.

[cardosofede]

Pls test that works @rapcmia btw i think that we should evaluate doing it async by default

[rapcmia]

Could you please review this PR? It fixes a blocking issue in get_active_containers.

Pls test that works @rapcmia btw i think that we should evaluate doing it async by default

I have a look later this afternoon 🙏🏼

[rapcmia]

PR update:

• Setup this PR on linux (ubuntu24.04) using source build for both dashboard and docker for broker (emqx)
• Setup successful and able to launch hummingbot-deploy ✅
  • Successfully add credentials, upload controller-config (pmm_simple) then deploy ✅
• Observed error on backend-api logs
  

  INFO:     Waiting for application startup.                                                                                                                    
  2025-03-14 20:19:28,243 - asyncio - ERROR - Task exception was never retrieved                                                                                
  future: <Task finished name='Task-14' coro=<BotsManager.update_active_bots() done, defined at /home/yawnyunehh/hummingbot/backend-api/59/services/bots_orchest
  rator.py:88> exception=TypeError("'coroutine' object is not iterable")>                                                                                       
  Traceback (most recent call last):                                                                                                                            
    File "/home/yawnyunehh/hummingbot/backend-api/59/services/bots_orchestrator.py", line 97, in update_active_bots                                             
      for bot in active_hbot_containers:                                                                                                                        
  TypeError: 'coroutine' object is not iterable                                                                                                                 
  /home/yawnyunehh/miniconda3/envs/backend-api/lib/python3.10/asyncio/runners.py:44: RuntimeWarning: coroutine 'BotsManager.get_active_containers' was never awa
  ited                                                                                                                                                          
    return loop.run_until_complete(main)                                                                                                                        
  RuntimeWarning: Enable tracemalloc to get the object allocation traceback
  

---

Test when no container active

• Deployed an instance with pmm simple controller on kucoin, no instance available even tho a container has been running for 4mins already. Orders are placed and cancelled
  
  • Deployed another instance and still not showing as active
  • Refresh browser and even backend-api still no go
  • Screen recording: https://www.loom.com/share/2f649e5a62554164992d90b2b291de22
  • Compare with latest backend-api ✅
    
    • Instance created ok

Steps to reproduce:

1. Setup source build for this PR and dashboard
2. Clone deploy then docker compose -up emqx for broker
3. After successfully uploaded a config, deploy an instance and observe backend-api logs
   • Make sure "POST /create-hummingbot-instance HTTP/1.1" is 200 ok
   • Observe deploy page and refresh if instance is created/active
   • Relaunc backend-api and check again
   • On terminal, run docker ps to check active containers
4. Observed that on instance page, no container is active ❌

[MarsYoung]

add await to sovle returns a coroutine for active containers

[MarsYoung]

PR update:

• Setup this PR on linux (ubuntu24.04) using source build for both dashboard and docker for broker (emqx)

• Setup successful and able to launch hummingbot-deploy ✅

  • Successfully add credentials, upload controller-config (pmm_simple) then deploy ✅

• Observed error on backend-api logs
  

  INFO:     Waiting for application startup.                                                                                                                    
  2025-03-14 20:19:28,243 - asyncio - ERROR - Task exception was never retrieved                                                                                
  future: <Task finished name='Task-14' coro=<BotsManager.update_active_bots() done, defined at /home/yawnyunehh/hummingbot/backend-api/59/services/bots_orchest
  rator.py:88> exception=TypeError("'coroutine' object is not iterable")>                                                                                       
  Traceback (most recent call last):                                                                                                                            
    File "/home/yawnyunehh/hummingbot/backend-api/59/services/bots_orchestrator.py", line 97, in update_active_bots                                             
      for bot in active_hbot_containers:                                                                                                                        
  TypeError: 'coroutine' object is not iterable                                                                                                                 
  /home/yawnyunehh/miniconda3/envs/backend-api/lib/python3.10/asyncio/runners.py:44: RuntimeWarning: coroutine 'BotsManager.get_active_containers' was never awa
  ited                                                                                                                                                          
    return loop.run_until_complete(main)                                                                                                                        
  RuntimeWarning: Enable tracemalloc to get the object allocation traceback
  

Test when no container active

• Deployed an instance with pmm simple controller on kucoin, no instance available even tho a container has been running for 4mins already. Orders are placed and cancelled
  

  • Deployed another instance and still not showing as active

  • Refresh browser and even backend-api still no go

  • Screen recording: https://www.loom.com/share/2f649e5a62554164992d90b2b291de22

  • Compare with latest backend-api ✅
    

    • Instance created ok

Steps to reproduce:

1. Setup source build for this PR and dashboard

2. Clone deploy then docker compose -up emqx for broker

3. After successfully uploaded a config, deploy an instance and observe backend-api logs

   • Make sure "POST /create-hummingbot-instance HTTP/1.1" is 200 ok
   • Observe deploy page and refresh if instance is created/active
   • Relaunc backend-api and check again
   • On terminal, run docker ps to check active containers

4. Observed that on instance page, no container is active ❌

this result was caused by the missing of await keyword,fixed it @rapcmia

[rapcmia]

• Setup this PR with latest dashboard and deploy(broker) on source
• Build docker image for tests ✅
• Reported issues has been fixed by developer ✅
• Add credentials and observe diff from development branch
  • Loads portfolio successfully

---

Test what happens when there are no containers?

• All ok, Launched 2 instances
• Relaunch backend-api and dashboard
  • 1 instance paused and other is active ✅
  • Start the instance and worked ok, 2 instance actively fetched from backend-api and dashboard/instance page

---

Testing with a mix of running and stopped containers

• Pause an instance while another is running then relaunch backend-api ✅
• Start an instance while another is paused observed behavior after backend api is launched ✅
• Paused all instance, relaunch backend-api then start all instance again ✅
• Archived all bots at the same time and observe behavior ✅

---

Run same tests cases with docker build ✅