StyleSpec Pro

Editorial · Dark · Technical

Nightshift.

A precise dark system for tools that need evidence, hierarchy, and calm under pressure.

Afterlog / Live incident field Production · us-east-1
LIVE / INC-0427

Checkout latency after deploy

14:32:08 UTC
Primary signal API response drift

Started 6 minutes after payments-api@4.18.0 reached 100% traffic.

deploy
14:1214:2214:32
$ afterlog diff inc-0427 --since deploy
findingconnection pool saturationpayments-api / db.primary

Nightshift Landing

Afterlog / Public site
AFTERLOG.
ProductField notesDocs
Open console ↗
RUNTIME INTELLIGENCE

Every incident is a sequence.

Afterlog gives engineering teams one field for the change, the signal, and the owner. Start with the event that moved production.

$afterlog inspect --service payments-api
  • 14:26:02deploy.complete4.18.0
  • 14:31:44latency.thresholdp95 1.84 s
  • 14:32:08incident.openedINC-0427
Incident field / 0427

The change and the consequence, on one axis.

No card wall. No summary detached from evidence. The timeline stays attached to the services it describes.

14:1214:2214:32
edge
checkout
payments
database
deploy / 4.18.0
OTLP in. A readable system out. Trace / Logs / Deploys / Ownership

Nightshift Product

Afterlog / Workspace
Incidents/INC-0427
⌘ KMT
Investigating

Checkout latency after deploy

Opened
14:32 UTC
Owner
Payments
Region
us-east-1
Event timelineLive / 12 events
  1. Deployment startedpayments-api@4.18.0release / maya.t
  2. Traffic reached 100%New instances healthydeploy / production
  3. Latency crossed thresholdp95 moved from 580 ms to 1.84 ssignal / checkout.latency
  4. Incident openedPayments team assignedafterlog / automatic
  5. Pool saturation isolateddb.primary / max connectionsfinding / query diff
$afterlog compare 4.17.3..4.18.0 --service payments-api

Nightshift Mobile

An incident handoff with only the active signal, latest finding, and next action.

AFTERLOGINC-0427
Investigating

Checkout latency after deploy

Owner
Payments
Opened
14:32 UTC
Latest finding / 14:34 Connection pool saturation

Retry behavior keeps connections open across failed payment intents.

  1. Latency crossed thresholdp95 1.84 s
  2. Incident openedPayments assigned
  3. Pool saturation isolateddb.primary
Open incident field ↗

Nightshift System

Actions and navigation

NS / 01
AFTERLOGIncidentsServicesDeploys
Open incident Compare deploys Dismiss
Live Degraded Failed

Command and input

NS / 02
⌘ F
Search by service, owner, or release.
$afterlog inspect --since 30m↵
!

Query stopped at line 12. The field span.owner is not available. Use service.owner.

Runtime stream

NS / 03
TimeEventSourceState
14:34:51pool.saturateddb.primaryopen
14:32:08incident.openedafterloglive
14:31:44latency.thresholdcheckoutlinked
Reading trace segments
No saved viewsPin a query to return to it here.
AGENT PACK / V0.1

Inside the pack

Five working files turn the visual direction into instructions an agent can apply beyond this showcase.

DESIGN.mdLAYOUT / COMPOSITION
Start with evidence, not a summary card. Put the change, signal, time, source, and owner on the same visual plane.
AGENTS.mdOPERATING RULE

Keep data and evidence attached to their source, time, owner, or state.

tokens.cssREAL VALUES
--color-paper: oklch(13% 0.012 145);
--color-accent: oklch(78% 0.155 145);
--font-display: "Space Grotesk", "Trebuchet MS", sans-serif;
REFERENCES.mdMOBILE / FINDING
<article class="mobile-finding">
  <span>Latest finding / 14:34</span>
  <strong>Connection pool saturation</strong>
</article>