trinitrotoluene / staytusdaemon Goto Github PK

In order to enable the daemon to be useful more broadly, it should support drop-in .dlls providing implementations for a shared interface of plug-in "downtime sources".

What exact API this ends up being is still TBC but the current MC and ICMP checkers will be moved to a separate assembly and loaded in the same way as a consequence.

The service unit directive type will need to be expanded upon with version information to ensure only compatible assemblies are loaded.

Once a service reaches a given degree of outage, an issue should automatically be created.

Requires

• New config options
• Template strings(?)
• Check Staytus API spec since creating issues is undocumented

Currently, the service is polled once each interval and then recategorised based on its current and previous statuses.
In order to make the daemon more versatile the following scenarios should be enabled:

• Configurable retry count before deciding a service is offline
• Latency threshold for "degraded-performance"
• Failure count threshold for
  • "partial-outage"
  • "major-outage"
• Success count threshold for
  • "operational"
  • FROM degraded/partial/major
• Optional: Export service latency to logfile for charting?

Once Docker base images become available for the 3.0 SDK, the service should be dockerized.

• HTTP
  • Simple GET
    • Configurable desired response status code, default 200 OK
  • WebSocket(?)
    • Check for 101 Switching Protocols
• Steam
  • Target game id