Drop In The Agent.
One line per service. REST or GraphQL. Spans start flowing the moment it boots.
$ wft install --workspace brightcask
00 / 07 · API monitoring + distributed tracing
Your Next 3 A.M. Page Names The Hop.
Trace every API call out of the labyrinth. REST. GraphQL. Every hop. One thread.
01 / 07 · On the rota
On-Call Teams Read Traces, Not Logs.
- Cinder Park
- Oakmont
- Tidewell
- Yarrow & Co
- Quillon
0probe regions
0days of trace retention, top plan
0click from alert to trace
02 / 07 · Features
Five Tools. One Thread.
Trace it. Probe it. Page on it. Budget it. Diff it.
FIG.01 — trace.waterfall · span flagged
FT-01 / distributed tracing
One Request. Every Service.
The gateway hands off to six services. We stitch each span into one waterfall. The slow hop lights up orange.
FT-02 / synthetic checks
Probe From 25 Regions. Hear It First.
Synthetic checks hit your REST and GraphQL endpoints from 25 regions. A bad region shows as one hot cell, not a vague customer email.
- 25 probes
- 1-minute checks from Spindle up
FT-03 / alerts with traces
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.
- PagerDuty + webhooks from Labyrinth
- Email + Slack on every plan
FT-04 / SLO tracking
Watch The Error Budget Drain.
Set latency and error-rate SLOs per endpoint. See the budget burn day by day. The deploy that bled it gets named.
- Latency + error-rate targets
- Custom SLO dashboards on Minotaur
FIG.05 — schema.diff · GET /v2/orders
FT-05 / schema drift
Drift Shows Up As A Diff.
REST or GraphQL, every response is checked against its last known shape. A number turned string lands with the deploy that did it.
Schema drift detection
03 / 07 · Setup
From Install To First Trace
Add the agent. Aim the checks. Wait for the page.
Aim The Checks.
Point probes at your endpoints. Set the SLO. Choose who gets paged and where.
check POST /v2/checkout every 1m slo p95 < 450ms → page #oncall-api
Sleep. Get Paged. Read.
The next alert shows up with its trace attached. You open one link and see the hop.
page → tr_8f2c19a hop postgres · 1,184 ms
04 / 07 · Integrations
Works With Your Favorite Tools
Alerts land where the team already is. Six hookups, no new tab.
- INT-01Jiraticket from alert
- INT-02Trellocard per incident
- INT-03Intercomstatus to support
- INT-04Calendlybook the review
- INT-05Mailchimpoutage notices
- INT-06Xeroplan invoices
05 / 07 · Plans
Pick Your Spool.
Three plans. Same probes, same traces. The tier sets how far the thread runs.
Billed annually · prices in USD
PLAN-01
Spindle
One thread. Start small.
$219/yr · billed annually
- Up to 10 monitored endpoints
- 1-minute uptime checks
- 7-day trace retention
- Email and Slack alerts
PLAN-02
Labyrinth
Tracing across services.
$449/yr · billed annually
- Up to 150 monitored endpoints
- Distributed tracing across services
- 30-day trace retention
- PagerDuty and webhook integrations
Most popular
Minotaur
No ceilings.
$699/yr · billed annually
- Unlimited endpoints
- Custom SLO dashboards
- 1-year trace retention
- SSO and audit logs
06 / 07 · Field notes
Logs Say It Broke. Traces Say Where.
Find the hop. Fix the hop. Go back to bed.
The alert showed up with the whole trace attached. I saw the slow Postgres lock before my laptop finished waking.
Schema drift caught a number turned string in our orders response. It landed as a diff, not a client bug report.
The SLO dashboards ended the argument about whether latency was fine. The budget bar answers it.
07 / 07 · Your move
Get The Map Before The Maze Gets You.
Pick a plan. Aim it at one endpoint. Tonight’s first slow request shows its whole route.
Get started