One alert per problem

Alerts inside FirstSales are supposed to tell you when something needs your attention. A set of fixes this cycle makes sure that's actually what they do: no more duplicate alerts for the same issue, alerts that clear once the problem behind them is actually resolved, and credit warnings that only go to accounts they're relevant to.

What changed

Three separate problems with the alerting system are fixed. First, alerts could be created more than once for the same underlying issue, so instead of seeing one alert about a problem, you might see several nearly identical ones stacking up, which made your alerts list noisier and harder to trust than it should have been. Alert creation is now handled so that the same issue produces one alert, not several.

Second, alerts that should have cleared once their underlying problem was fixed sometimes didn't. An alert about, for example, a domain authentication issue could remain marked as active even after the actual problem behind it had been resolved, which meant you couldn't trust that an open alert in your list still reflected something you actually needed to act on. Auto-resolve logic is now corrected so alerts that should clear, do clear, once the condition that triggered them is no longer true.

Third, low-credit warning alerts were reaching accounts they shouldn't have, including accounts that weren't active or on a trial. That's scoped correctly now: AI-credit-exhaustion alerts only go to accounts that are actively using credits, active subscriptions and trials, rather than firing for accounts where a low-credit warning wouldn't mean anything actionable.

The duplicate-alert problem and the stuck-open problem often compounded each other in practice. If the same issue generated more than one alert, and then none of those duplicates reliably cleared once the issue was fixed, your alerts list could end up with several stale entries all pointing at a problem that no longer existed, all still sitting there as if they needed attention. Fixing alert creation to produce one alert per issue, and fixing resolution so that alert actually clears when it should, addresses both halves of that at once rather than just trimming duplicates without also fixing whether they go away.

The domain-authentication watchdog that generates some of these alerts has also been hardened as part of this work, specifically around how it detects and reports on authentication problems, so alerts tied to domain-auth issues are both less likely to duplicate and more likely to resolve cleanly once the underlying DNS problem is actually fixed. That pairs directly with the sending-domain verification improvements shipped this same period: a domain that's properly verified shouldn't be generating an open alert that never clears.

For the credit-warning scoping specifically, the practical effect is that these warnings now target the accounts where a warning is actually something to act on. An active or trial account approaching its credit limit genuinely needs to know before it runs out and campaigns stop sending; an account outside that state doesn't need the same interruption, since a low-credit warning wouldn't correspond to anything they could act on in the moment.

How to use it

There's nothing to configure for any of these fixes; they apply automatically to how alerts are created and resolved across your account. If you check your alerts list regularly, you should notice fewer duplicate entries for the same underlying issue, and you should be able to trust that an alert still showing as open reflects something that's still actually a problem, not something that was already fixed.

If you're on an active subscription or trial and using AI credits, you'll still get low-credit warnings as your usage approaches your limit, the way you would expect. If your account isn't active or on trial, you won't be interrupted with a credit warning that doesn't apply to your situation.

None of these fixes change what triggers an alert in the first place, a domain authentication problem, a credit threshold, or another condition FirstSales is already watching for still generates an alert the way it always has. What changes is the reliability of what happens after that: one alert instead of several, a clean resolution once the issue is actually fixed, and delivery scoped to the accounts a given alert type is actually relevant to.

Why it matters

An alert system only works if you can trust it. If the same problem shows up as three separate alerts, you either waste time confirming they're all about the same thing, or worse, you start tuning out alerts altogether because your list is cluttered with duplicates. If an alert stays open after its underlying issue is already fixed, you can't rely on your alerts list to tell you what still actually needs attention, which defeats the purpose of having alerts at all.

The credit-warning scoping fix is smaller but still matters for the same reason: a warning that doesn't apply to your account's situation is just noise, and noise is what makes people stop paying attention to alerts that do matter. Cleaning all three of these up means when you see an alert in FirstSales now, it's there because something genuinely needs your attention, once, and it goes away once that's no longer true.

If you manage alerts for more than one workspace or client account, this consistency matters even more: a clean, accurate alerts list is the difference between quickly scanning for what actually needs attention across several accounts and having to manually verify whether each open alert is still real before you can trust your own summary of what's going on.

A short, accurate alerts list is worth more than a long one, and that's the outcome these three fixes are aimed at together: fewer alerts, each one real, each one clearing when it's actually resolved.

Expect your alerts list to look noticeably shorter after this ships, not because fewer real problems are being surfaced, but because the noise sitting on top of them is gone.

Availability

These fixes are live for all accounts now and apply automatically to alert creation, resolution, and credit-warning delivery going forward.