About Agent Persistence

[ianido]

Hello, I started using your package to connect to a centralized Hangfire Server by http, however I do not understand the concept of Agent and their Persistence (mssql, redis and mysql), Why Agent needs direct access to the hangfire database? What is the difference between having an agent and simply using the hangfire clases?

[yuzd]

Agent Can be disassembled into different applications according to business lines
In this way, different business lines can write jobs in their own applications

Why Agent needs direct access to the hangfire database?
It is because the status and execution log of the job need to be synchronized to hangfireserver
For example, unified management in hangfire dashboard

By the way
It may be better to switch to communication between agent and hangfireserver
How do you think?

[ianido]

Thank you for your answer, I am still not sure what is the difference in between having an Agent in a windows console and having a windows console with hangfire server, I can have as many servers as I want, I can define a server looking at specific queue as well.

Not sure which scenario is covered by Agents specifically than a regular hangfire server app can't cover?

[yuzd]

From a design perspective, I think the Hangfire.httpjob,agent interacting directly with storage is a bad design,
and it would be better to change it to interact with hangfireserver.
At first, I wanted to make the agent's status and logs unified and manageable in hangfire.dashbord.
At that time, I just designed it like this for convenience.

[ianido]

100% agrees

[maxiar]

Hi! any Plan to improbe this to isolate the hangfire database from agents ? .. :) I can help us if you think any design..