Component status updated based on the incidents

Question

Hi!

I have and issue that seems to be something many people before probably have struggled:

I have an integration in opsgenine that creates statuspage incidents. Incidents are assinged to components (I use these opsgenie tags like cmp_myapp:degraded_performance). There is a possibility for multiple incidents for same component. The question is:

Is there a way to make component change status on its own to "operational" when all the incidents are resolved, without any API interaction?

Using the opsgenie integration I can either not update the status of a component at all (and when I close all the incidents the component will be still not operational) or it will change the status of the component to operational on first alert closed, no matter if there are other alerts(incidents) related to this component.

Kind Regards,

Przemek

nico kruger · Answer

Can you expand on this please @Ronnie Kingston ? We are facing the exact same issue when both alert A and alert B from opsgenie is changing the component status, but then B temporarily resolves (let's say A means a database is slow, and B means a request succeeded. even though the database is still slow and 99% of all requests are failing every now and again one comes through which makes B resolve) and changes the status to operational.

Any guidance would be really helpful on this, thanks.

Ronnie Kingston_snohup · Answer

Hi Przemek,

We have a configuration saying "When an alert is closed in Opsgenie, Resolve the incident in Statuspage and update the component statuses according to the Opsgenie tags".

Every incident in Opsgenie will have an owner alert. When the incident is closed in Opsgenie the owner alert is also closed. This alert can be used to communicate incident statuses to Statuspage via the integration.

Having said the above, if there are multiple un-related incidents in Opsgenie impacting the same component in Statuspage, then, when you close the first incident in Opsgenie, that will close the owner alert and therefore will also mark the Statuspage component as operational even though there are other open incidents in Opsgenie.