Is your feature request related to a problem? Please describe.
Being able to execute commands with an endpoint and without authentication makes the agent a vulnerability in itself.

Describe the solution you'd like
Implement auth, at least with the token provided in the config file. Use user token, or agent token.

Describe the solution you'd like
Receive POST with command as string and return a JSON with keys: stdout, stderr and elapsed time.

Describe the solution you'd like

• Serve it as package in PyPi.
• Provide a Windows Service and Linux Service file to ease the installation.

• Extract metrics
• Format and send metrics
• #2
• #3
• #13
• #14
• #4
• #24
• #11
• #6
• #12
• #15

Is your feature request related to a problem? Please describe.
Logging is managed with Open() instead with Logging module.

Describe the solution you'd like
Change logging to Logging module.

Split config file in sections.

{
    "uvicorn": {
        "host":"0.0.0.0",
        "port":8000,
        "workers":4,
        "reload":true,
        "debug":true,
        "log_level":"trace",
        "ssl_keyfile":null,
        "ssl_keyfile_password":null,
        "ssl_certfile":"None",
        "ssl_version":3,
        "ssl_cert_reqs":null,
        "ssl_ca_certs":null,
        "ssl_ciphers":"TLSv1",
        "limit_concurrency":null,
        "limit_max_requests":null,
        "backlog":2048,
        "timeout_keep_alive":5
    },
    "metrics": {
        "endpoint":true,
        "url":"http://httpbin.org/post",
        "post_interval":60,
        "enable_file":false,
        "metrics_file":"metrics.json"
    },
    "alerts":{
        "url":""
    }
}

Describe the solution you'd like
Document with Google Style the code when the agent is done, so it's done when finishing the Final Degree Project to directly create an HTML with Sphinx.

Optimize and refactor the code

Describe the solution you'd like
User may don't want all of the data from the host. Use templates to select which data is required to be processed.

Is your feature request related to a problem? Please describe.
When receiving the JSON, the message doesn't have any breaklines and makes it illegible.

Describe the solution you'd like
Replace \n with some characters and replace it again for a
in the Web app.

Describe the bug
Adjust Disk Metric according to this issue.

Describe the solution you'd like
Extract the number of processes and processes that surpass a configurable threshold in performance.

Is your feature request related to a problem? Please describe.
Numeric metrics like space (bytes, ...) and time are strings but will cause problem in the server in terms of:

• Storage due to encoding
• Unformatting of every metric of every host in each petition.

Describe the solution you'd like
Not format the metrics and leave that load to the web client.

Describe the bug
Executing an interactive command blocks incoming requests until it finishes.

To Reproduce
Steps to reproduce the behavior:

1. POST: localhost:8000/command?command=python -m http.server 8080&timeout=60
2. POST: localhost:8000/command?command=dir&timeout=5

Expected behavior
First POST launch and be in background or at least not blocking second POST.

Desktop (please complete the following information):

• OS: Windows
• Version 1.0.0

Additional context
Check Subprocess calls

Is your feature request related to a problem? Please describe.
Does Uvicorn logs into a file?

Describe the solution you'd like
If not, use Gunicorn and use it's logging mechanism.

Describe the solution you'd like
Implement an Alert Manager in each Agent to trigger when a specified value surpass the value set in a JSON config file.
This Alert Manager may send the request with a POST or a Webhook.

Additional context
Implemented in the Agent to remove the load from the REST API server (many values for many hosts to check, may need to be implemented with Celery but this is more flexible and lightweight).

Note: This doesn't remove the need of an alert system in the REST API server for when a hosts is down.

Is your feature request related to a problem? Please describe.
Agent needs to listen for POST data (command or new settings) but as of now it's not possible because of the execution flow.

Describe the solution you'd like

• Launch 2 threads and set Timeloop and FastAPI in each of them.
• Launch Timeloop from FastAPI.

Is your feature request related to a problem? Please describe.
It improves flexibility in the Agent execution flow because it allows not to interrupt the program manually and launch it again.

Describe the solution you'd like
Receive settings as JSON. Replace Pydantic for a new designed class.
When new settings are received, refresh the server with the new settings. Authenticate with the agent with a token, else return auth error.

Describe the bug
Agent does not test type and data of the parameters from the settings file.

To Reproduce
I can POST this invalid JSON config without any error showing up:

{
    "alert_url": "",
    "backlog": 2048,
    "debug": "a",
    "host": "300",
    "limit_concurrency": "None",
    "limit_max_requests": null,
    "log_level": "trace",
    "metric_enable_file": false,
    "metric_endpoint": true,
    "metric_file": "metrics.json",
    "port": 8000,
    "post_interval": "a",
    "post_metric_url": "http://httpbin.org/post",
    "reload": true,
    "ssl_ca_certs": null,
    "ssl_cert_reqs": null,
    "ssl_certfile": null,
    "ssl_ciphers": "TLSv1",
    "ssl_keyfile": null,
    "ssl_keyfile_password": null,
    "ssl_version": 3,
    "timeout_keep_alive": 5,
    "workers": 4
}

Expected behavior
It should return an error for each of the invalid parameters.