Wait X minutes before next check?

[b4bbuilder]

I want to set a number of watches at 5-minute intervals but make it so that after one is finished checking another cannot be run for a specified delay. For example, if watch #1 is currently processing and watch #2 is in queue, I would like to wait a minute after watch #1 is finished before watch #2 is processed. Is this currently possible?

I did find the environment variable MAX_QUEUE_LENGTH which I could set to 1 but that will result in a longer delay that I would like.

[robbedenys]

Hi, I also would like to limit the number of watches being checked simultaneously. At the moment it runs 10 watches at a time, but I cannot find where to set this limit? I tried the MAX_QUEUE_LENGTH (set it to 1 and to 100), but that does not make a difference. I also changed the "MAX_CONCURRENT_SESSIONS" playwright-chrome setting to 1, but it still performs 10 checks at once.

I would love to have the watches checked every 15 minutes, but with only one watch being checked at the same time. Is this possible in any way?

[robbedenys]

Still haven't found a way to do this. The problem is that when there are too many concurrent connections to the website I am monitoring, access is blocked for some time. There is no problem when all the watches are run in immediate succession.

I now manually start the watches (fortunately no more than 20) and wait for the previous watch to finish before starting another one. But if I need to restart my server for some reason, I have to do this manual process again. I hope there is a setting that solves this.

[Giattuck]

Hello @robbedenys have you find a solution? I'm searching to solve the same exact problem

[robbedenys]

Yes, I have found a solution that works quite good. I am using a home assistant automation that runs all the watches one after another.

I created this python script that runs within home assistant:

#!/usr/bin/python3
import logging
import requests
import json
import asyncio

_LOGGER = logging.getLogger(__name__)

def run_request(header, url):
    # Use task.executor to run the request function
    response = task.executor(make_request, url, header)

    # Extract response content or status code as needed
    response_text = response.text
    response_status_code = response.status_code

    # _LOGGER.error("Response Text: %s", response_text)
    if response_status_code == 200:
        return response_text

@pyscript_compile
def make_request(url, headers):
    try:
        response = requests.get(url, headers=headers)
        return response
    except requests.exceptions.RequestException as e:
        # Handle any request exceptions (e.g., network issues)
        print(f"Request Exception: {e}")
        return None

@service
def run_watches(watches_count):
    header = {"x-api-key": 'YOURAPIKEY'}
    url = 'http://IPANDPORT/api/v1/watch?tag=enabled'

    data = run_request(header, url)
    parsed_data = json.loads(data)
    # _LOGGER.error(len(parsed_data))
    count = int(len(parsed_data))
    service.call("input_number", "set_value", entity_id=watches_count, value=count)

    for item_id, item_data in parsed_data.items():
        url = f'http://IPANDPORT/api/v1/watch/{item_id}?recheck=1'
        data = run_request(header, url)
        # _LOGGER.error(data)
        asyncio.sleep(60)

you should replace YOURAPIKEY and IPANDPORT with your own API key and the IP and PORT of your own installation.

Then in home assistant I made a numeric entity that keeps track of the total number of watches that has a specific tag (in my case "enabled") that gets filled in by the python script.

This is the yaml code of my automation:

alias: Changedetection
description: ""
mode: queued
max: 10
triggers:
  - value_template: >-
      {% set watchcount = states('input_number.changedetection_watches_count') |
      int%}   

      {{ (now().timestamp() | int % (watchcount*60)) == 0 }}
    trigger: template
conditions: []
actions:
  - data:
      watches_count: input_number.changedetection_watches_count
    action: pyscript.run_watches

All my changedetection watches are paused and i force check them with the python script. It works fine for more than a year already :-) When you add a watch the next time the automation is triggered it will automatically also check it, so once it is running you don't have to think about it anymore..

[Giattuck]

@robbedenys very interesting, thanks! How many links do you have on a single website? because i have around 6500

[robbedenys]

Oh wow, I have around 50..
You could group them with different tags and have the automation trigger the script multiple times simultaniously..

[ta7tUHIJHx4RwR21]

change - FETCH_WORKERS=10 to - FETCH_WORKERS=1 in docker-compose.yml to make change detection check one at a time.