We would be interested in a service/add on that would allow us to subscribe to metrics similar to aws cloudwatch where parameters are set to receive email/webhook notifications when metrics start to fall outside the status quo.
I understand this can be achieved with integrations with the prometheus api - but being able to add simple alerts from the dashboard would be a huge win.
Ahh this is great, we will get some of these setup and report back on how they go.
I have no idea if this is at all something that fly wants within its scope, but setting up simple alerts for metrics like response time in the fly dashboard would be super nice
I’m having a little trouble with this. I’m on https://fly-metrics.net/ in my org (of which I am owner, and my account says I am an Editor on fly-metrics.net). I can Explore and create a new Dashboard, but the links above talk about “1. In the left-side menu, click Alerts & IRM and then Alerting.” and I don’t have that in my sidebar.
I’m currently on a Hobby plan, but I’m likely ready for the Launch plan. Will that get me alerting?
I think kurt meant if you are self-hosting a full-fledged Grafana instance on fly. Afaik the fly-metrics Grafana has several limitations, one of them being not being able to create alerts.
+1 there’s no option right now to email/page alert for failing healthchecks and other key metrics like memory pressure. will be willing to pay for this and it would allow us to move off of third-party monitoring providers that are only aware of app-level issues and not infrastructure-level like fly.io would provide. we had downtime recently that went unnoticed because of high memory pressure and when doing postmortem i stumbled here seeing there’s no clear option to setup infra alerts
We’re also looking for an alerting solution and had hoped to leverage the Grafana instance that Fly provides with each app. We understand that full alerting functionality might not come for free and would be open to paying for it—but the lack of a path forward puts us in a tricky spot.
Right now, we’re facing a choice between duplicating metrics and logs to another provider just to get alerts, or continuing to use Fly’s Grafana without the ability to act on the metrics we’re already capturing.
Is there any update or roadmap from the Fly team on whether alerts will be supported in the future? Even a paid or limited version would help us plan better.
Adding alerting to the existing Fly hosted Grafana instance is not planned for the near future.
The simplest option is to set up a self-hosted Grafana instance within your Fly org and use that for setting up alerting. It can read from the same metrics stream as the Fly-hosted version, but since it’s a full standalone Grafana install you can set up alerts and modify it as you like.
It’s quick to set up, and should only cost a couple of dollars a month to run.