Filter By Hop
Show only requests where postgres took over 500 ms. Method, status, service, deploy.
FEATURES · spec sheet · 7 entries
Next Incident, You Open One Link. Hop Found.
Five tools. Seven checks on the stack. Each one shortens the walk from page to fix.
FT-01 / distributed tracing
One Request. Every Service. One Waterfall.
The Slow Hop Lights Up.
The gateway hands off to services. We stitch each span into one tree. REST or GraphQL, the longest wait turns orange.
- Spans nested by caller
- Deploy hash on every trace
- Step through with J and K
Workflow fit
Alert fires. You land here. You read the lock wait. You never open a log search.
FT-02 / synthetic probes
25 Regions. One Hot Cell.
Hear It Before Customers Do.
Probes hit your endpoints every minute from 25 regions. One bad region shows as one red cell with a trace captured, not a vague email.
- Three slow runs trigger the page
- Trace captured on every failure
- REST and GraphQL bodies
Workflow fit
Regional trouble stays regional. You see the cell before the dashboard averages it away.
FT-03 / trace-attached alerts
The Page Carries The Failing Trace.
No dashboard hunt at 3 A.M. The alert arrives with the trace, the slow span and the deploy before it.
Open It. Read It. Fix It.
Every alert names the slow span and the deploy before it. Responders see the route, not a threshold number.
- PagerDuty, Slack, email, webhooks
- Acknowledge with one key
- Mute for 1 h from the page
Workflow fit
Page, trace, ticket. Link the incident into Jira and the card carries the same trace.
FT-04 / SLO budgets
Watch The Error Budget Drain.
The Deploy That Bled It Gets Named.
Set latency and error-rate targets per endpoint. See the budget burn day by day. A spike links to the commit behind it.
- Rolling 30-day windows
- Burn alerts before the budget is gone
- Custom dashboards on Minotaur
Workflow fit
Friday release review opens on this panel. The argument about “is latency fine” ends here.
FT-05 / schema drift
Drift Shows Up As A Diff.
A Number Turned String. Caught.
Each response is checked against its last known shape. Breaking and additive changes are labeled. The deploy behind them is named.
- REST JSON and GraphQL
- Breaking vs additive
Workflow fit
Clients break quietly. This breaks loudly, on your side, before the bug report.
FT-06 / request log
Every Slow Request. One Click To Its Trace.
Bulk Act
Select rows. Open every trace, mute for an hour, assign an owner.
Retention By Plan
7 days on Spindle. 30 on Labyrinth. 365 on Minotaur.
FT-07 / command search
Press ⌘K. Land On The Trace.
Search endpoints, trace ids, deploys and teammates from anywhere. Keyboard first. No menus to dig through.
Workflow fit
Paste a trace id from a support ticket. You are on the waterfall in two keystrokes.
SPEC / security + reliability
What Runs Underneath.
Read the table. Skip the pitch.
SEC-01
Access control
SSO and audit logs on Minotaur. Every acknowledge, mute and setting change is recorded with who and when.
SEC-02
Data in the trace
Scrub rules strip headers and body fields before a span leaves your service. Card numbers and tokens never reach a trace.
REL-01
Probes that watch the watcher
25 regions check each other. A dead probe region is flagged, not counted as a pass.
REL-02
Agent overhead
The agent samples and batches spans. If it cannot reach us, it drops spans, never your requests.
SPEC / integrations
Alerts Land Where The Team Already Is.
Six hookups. No new tab.
- INT-01Jira
A ticket per alert, trace link inside.
- INT-02Trello
One card per incident.
- INT-03Intercom
Status posted to support.
- INT-04Calendly
Book the post-incident review.
- INT-05Mailchimp
Outage notices to customers.
- INT-06Xero
Plan invoices, synced.
END / your move
Seven Features. One Thread Out.
Pick a plan. Aim it at one endpoint. The next slow request shows its whole route.
Get started