Why This Alert

Why This Alert: Every Incident Explains Itself

An alert that only says "down" leaves the first ten minutes of the incident to guesswork. SutramX records the evidence and the decision behind every incident and puts a one-line verdict in the alert itself, including when the cause is probably not yours.

In short

Every SutramX incident has a "Why this alert" panel: each region's result (status, error class, HTTP status, timings), the quorum rule applied, the failure class in plain words, how many regions it affected, whether an alert was sent or why not, and a one-line verdict that is also included in the alert. It is rule-based, not AI, and on every plan including Free.

Deterministic, not generated

The explanation is built from the check results and the alerting rules that actually ran, with fixed rules. No language model writes it, so the same evidence always gives the same explanation, and it is available on every plan, Free included.

It answers the questions every responder asks first: did more than one place see this, what kind of failure is it, and why did (or didn't) my phone go off?

ProbefiresCheckfailsQuorumconfirmsAlertdispatches

The verdict travels with the alert

The one-line verdict is added to down alerts on email, Slack, Microsoft Teams, Discord, Google Chat, Mattermost and Telegram, to the details sent to PagerDuty and Opsgenie, and to webhooks as an `explanation` field. SMS, WhatsApp and voice alerts stay short and do not include it.

  • YoursEnough regions confirmed a failure at your endpoint itself: DNS, TLS, connection, timeout, an HTTP error or a content check.
  • ExternalA vendor your service depends on is failing for several SutramX customers at the same time.
  • CheckerThe evidence points at a SutramX checker rather than your service ("Our checker, not your site").
  • UnknownThe evidence does not point clearly anywhere; the panel still shows everything that was seen.

Flakiness score

Each monitor gets a flakiness score from 0 to 100 over the last 7 and 30 days, with up to three top reasons: failures that were never confirmed, short incidents that recovered on their own, flapping and inconclusive checks. A flaky monitor is worth tuning before its next alert is ignored.

Specifications

PlansEvery plan, including Free
MethodFixed rules over check results and alert decisions; no AI
Per-region detailStatus, error class, HTTP status, timings
Fault verdictsYours, external, checker, unknown
Verdict in alertsEmail, chat apps, PagerDuty/Opsgenie details, webhook `explanation` field (not SMS, WhatsApp or voice)
Flakiness score0–100 per monitor, 7 and 30 days, with top reasons

Learn more

Related capabilities

Vendor Outage DetectionKnow when it is Stripe, OpenAI or AWS, not you. SutramX spots vendor API outages from its own checks and compares them with the official status page.Explore Vendor Outage Detection Multi-Region ProbesSynthetic probes from 3 regions with multi-region confirmation on paid plans, catching regional routing failures and CDN misconfiguration.Explore Multi-Region Probes Alerting & EscalationEmail, Slack, Discord, Telegram, Teams, Google Chat and Mattermost alerts on every plan, plus webhooks, GitHub, PagerDuty and Opsgenie, with escalation.Explore Alerting & Escalation

Know it’s down before your customers do.

Start free — Free plan forever, no card required. Upgrade any time.

Start free See pricing

We'd like to use analytics cookies (Google Analytics and PostHog) to understand how sutramx.com and the SutramX app are used. They are only set if you accept. See our Cookie Policy.