The on-call desk that opens the fix.
Nightglass sits across your repos, staging, and production around the clock. When an incident fires it correlates the signals, scores the SLO, opens a fix PR with the rollback pre-wired, and ships the postmortem into the channel you already use.
- 02:14:08Z[detect] error_budget_burn = 4.2× on payment.charge
- 02:14:11Z[correlate] commit a91f3c · feat/checkout-retry-with-jitter · 38m ago
- 02:14:14Z[triage] SLO at 38%; route → #sre-checkout (Joon, on call)
- 02:14:21Z[open] PR #1841 opened · rollback wired · Slack draft staged
--- a/services/payment/retry.go
+++ b/services/payment/retry.go
@@ -42,7 +42,7 @@
- jitter = rand.Duration(0, 5*time.Second)
+ jitter = rand.Duration(0, 750*time.Millisecond)
if attempt >= maxAttempts { return errGiveUp }Four jobs the on-call desk already owns.
Nightglass is not another alert-correlation layer. It is a colleague who never sleeps, takes the fix end-to-end, and leaves the team with a clean commit, a clean channel, and a clean postmortem.
Nightglass fuses logs, traces, deploys, and the last hour of commits on the owning service the moment a SLO breach fires. Severity is judged against your error budget, not a generic threshold, so the alert you receive is the alert the on-call can act on.
Severity, service ownership, and recent incident history decide routing. Nightglass hands the page to the channel the owning team already lives in — PagerDuty, Opsgenie, Slack, or Teams — and keeps a record of who answered first so the audit trail stays clean.
Instead of waking an engineer to write the patch, Nightglass drafts a PR against the owning repo, embeds the rollback plan in the change description, queues a Slack draft for the on-call who owns the service, and stages a one-click revert behind the merge button.
When the SLO recovers, Nightglass auto-publishes a chat-native postmortem into the same channel where the incident lived — diff, timeline, customer-impact window, and the rollback steps your team would actually run. Nothing to chase; the record is already written.
From a fired SLO to a merged revert in five steps.
- 01
Detect
SLO breach fires across logs, traces, and deploy markers.
- 02
Correlate
Last hour of commits on the owning service is matched to the failing path.
- 03
Triage
Severity is scored against your budget; routing picks the destination.
- 04
Open fix
A PR is drafted with a rollback plan pre-wired and a Slack thread queued.
- 05
Revert ready
One-click revert sits behind the merge button; nothing the on-call has to write.
Built for the teams whose auditors already have opinions.
Nightglass runs against your code and your data, which puts it inside your change pipeline. Every action is logged, every prompt is reviewable, every prompt result is diffable — and the deployment is configurable to the boundaries your compliance team already maintains.
Audited continuously; reports available under NDA.
Audited continuously; reports available under NDA.
Audited continuously; reports available under NDA.
Audited continuously; reports available under NDA.
A good fit if your night reads like this.
- 01Platform & infra teams running dozens of services under one pager
- 02SRE groups who already trust PagerDuty-class alerting and want the next step
- 03Series B–enterprise SaaS, fintech, and e-commerce with revenue-bound SLOs
Bring Nightglass into a service that already pages you tonight.
Reach the team that builds it. One inbox; an engineer reads every message.
Join the waitlist.
One row. Drop your work email and the company you keep the pager for.
Join the Nightglass listOr email nightglass-3@polsia.app