Okay, I have to say I was genuinely pleased to see the new budget alerts feature hit my PromptLayer dashboard this morning. Setting spending thresholds and getting an email or Slack notification before a project spirals is such a foundational need for any team running LLM apps in production. It feels like a real step towards maturity for the platform.
But here’s my immediate reaction, and maybe others in the community feel the same: why stop at email and Slack? For those of us in B2B SaaS, especially where reliability and uptime are directly tied to LLM costs, this kind of alert needs to escalate into our existing incident response workflows.
The absence of a PagerDuty (or even Opsgenie) integration feels like a missed opportunity. If a budget alert fires, it’s often a critical signal—it could mean a runaway process, unexpected usage spikes, or something that needs immediate engineering attention. Having to manually bridge that gap by forwarding an email into PagerDuty creates a fragile link and adds toil.
I’m curious: is this on the roadmap? For teams practicing product-led growth, understanding cost drivers is crucial, but so is integrating those insights seamlessly into the operational tools we already use. A webhook option would also be a great middle ground, allowing us to build our own connections.
— Jane
— Jane
Totally feel you on the >escalate into our existing incident response workflows.
In my last role, we had our Salesforce dashboards wired to PagerDuty via webhook for exactly this kind of revenue-critical event. It's not just about getting the alert, it's about forcing the right person to acknowledge it immediately. When a cost spike is tied to a customer-facing feature, the support and engineering teams need to be looped in instantly, not whenever someone checks their email.
I'd bet this is coming soon. They probably rolled out the core email/Slack alerts first to get the feature live. PagerDuty integration likely needs a bit more engineering for their webhook system. Fingers crossed it's in the next release cycle!
> why stop at email and Slack?
You're not wrong. For any team running production LLM apps, a budget alert that doesn't trigger a war room is half a solution.
I'd push back a little though. PagerDuty is great, but not everyone uses it. Opsgenie, VictorOps, even a generic webhook to a custom endpoint. What I'd rather see is a simple webhook URL field in the alert settings. Let me wire it to whatever I want. That's less engineering for them and more flexibility for us.
Have they said anything about adding a webhook option? That would solve the PagerDuty gap for anyone who can build a quick integration.
Completely agree on the escalation point. We see this with Asana integrations too, where a simple Slack alert isn't enough for a critical-path blocker.
The webhook idea user773 mentioned is perfect - it would cover PagerDuty and let us build custom logic. Maybe it could even fire to a Zapier endpoint, and then you've got the whole connector universe.
But the real question is, what's the severity level? A budget alert could be "check this tomorrow" or "all hands on deck right now." I'd want to set that per-alert before it hits an on-call system.
Yeah, the PagerDuty gap is a big one. I've seen teams where the Slack channel for cost alerts just becomes background noise, and the real critical one gets missed because someone's phone is on silent. That manual email forward is a real toil multiplier.
On the roadmap question, I haven't seen anything official. But based on how they've rolled out other features, I'd guess a generic webhook comes first, then maybe specific integrations. That would let you plug into PagerDuty, Opsgenie, or even a small Lambda function to add custom logic before escalation.
What's your threshold set at? I'm curious if teams are using these for "watch this" alerts around 80% of budget, or only for the hard "stop everything" 100% breaches.
Pipeline Pilot