You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For the Golem Stats page we're scanning the relay's list of nodes and querying each individual one to confirm whether or not they are still online.
This is done in intervals of 30 seconds currently because we want as accurate a measure as possible with minimal risk of not spotting if a node goes down for a few seconds and then connects back again.
The main issue here is that it's a "best possible scenario" rather than a foolproof solution. If the relay allowed us to listen for events (SSE) about changes in connections from nodes, then we would be able to instantly identify these changes and display them on our stats page and that would provide a very great experience for our users.
Another reason why this is a great idea is that instead of having the stats page + reputation system send thousands of requests to confirm if a node is still online, then it could simply just listen and monitor for changes in the relay. Instead of producing thousands of requests, then it would probably just be a few requests a minute given that 70% of our network has an uptime between 90-100%.
Furthermore the stats page would serve as a solution to a users request like this one below because it can immediately see this.
The text was updated successfully, but these errors were encountered:
For the Golem Stats page we're scanning the relay's list of nodes and querying each individual one to confirm whether or not they are still online.
This is done in intervals of 30 seconds currently because we want as accurate a measure as possible with minimal risk of not spotting if a node goes down for a few seconds and then connects back again.
The main issue here is that it's a "best possible scenario" rather than a foolproof solution. If the relay allowed us to listen for events (SSE) about changes in connections from nodes, then we would be able to instantly identify these changes and display them on our stats page and that would provide a very great experience for our users.
Another reason why this is a great idea is that instead of having the stats page + reputation system send thousands of requests to confirm if a node is still online, then it could simply just listen and monitor for changes in the relay. Instead of producing thousands of requests, then it would probably just be a few requests a minute given that 70% of our network has an uptime between 90-100%.
Furthermore the stats page would serve as a solution to a users request like this one below because it can immediately see this.

The text was updated successfully, but these errors were encountered: