Cost spike
Statistical anomaly vs scope-aware baseline. Sustained or immediate.
Anomaly vs baseline band
+312% spike
7-day median · per hour-of-day
AI-Powered
Smart AlertingCatch cost issues before they hit the bill
Findthespike.Beforefinancedoes.
Statistical baselines, budget burn-rate, forecasted spend — across every workload, team, namespace. When spend drifts, the right owner hears about it.
- Four alert types
- Zero tuning
- 30-second root cause
Cost trend · payments-prod · 7 days
Firing · 2m 14s+312% spike
Sat · 12:48 UTC
Baseline
$24 /hr
Observed
$92 /hr
Excess (12h)
+$0
Detected in
0s
Routing · alerts → policies → channels
One alert. Rightinbox.
Slack, email, PagerDuty, webhook. Every alert flows through a reusable policy — caps, quiet hours, severity routing.
- Reusable policiesOne policy, many rules. Share across teams.
- Auditable end-to-endEvery attempt logged. Delivered, deferred, suppressed.
Alerts
Policy
Channels
Cost spike · ml-prod
+$1,8402m
Budget burn · data-platform
78% of $12K18m
Forecast · checkout-api
+$420 MTD1h
Policy
payments-prod
- modeimmediate
- cap10 / day
- quiet22:00–06:00 IST
- severityP1 · P2 · P3
Slack
#finops-alerts
finops@kubeadapt.io
PagerDuty
on-call · platform
Webhook
jira.example.com/hook
Last 24h · 14 / 14 deliveredimmediate · digest · in-app
Root cause · in 30 seconds
Every alert tells you WHY. NotjustTHAT.
Eight analyzers, parallel, against real Kubernetes events. Headline, recommended action, the workload that moved the needle — under 30 seconds.
- AI explanationsPlain-English summary and recommended action. Generated per incident, in seconds.
- Causal coverageWe tell you what we can't explain. Auditable to real K8s events.
- Typed causesReplica change. Image upgrade. HPA bounds. Specific events, not patterns.
8 analyzers · 30-second budget
Live incident
ml-prod · cost_spikeCost spike on ml-prod · checkout-worker driving 71% of delta
+$620/day+312% vs 14d baselineFirst time in 30 days
Top contributors
checkout-worker
CPU request changed
+$220/day
35%
ml-inference
Container image upgraded
+$180/day
29%
feature-store
Replica count changed
+$80/day
13%
Fully attributed. $480 of $620 delta explained by K8s changes.
Four signals · zero tuning
Four signals. Zerotuning.
Cost spike. Budget. New workload. Idle resources. Scope-aware baselines tune themselves.
One rule per scope · five scopes · all four signals
- Organization
- Cluster
- Namespace
- Team
- Workload
Budget threshold
Period budgets. Stay tuned.
Monthly burn
78%
day 23 / 31
50%80% warn100% capNew expensive workload
Flagged on first stable appearance.
First-seen stable
flagged
checkout-cache
namespace · payments
$1,240
/mo · NEW
Auto-excludes Argo · Spark · Ray · Kubeflow
Unused resources
v1.1Weekly digest of idle waste. Top-N by impact.
Weekly idle digest
top-N waste
- legacy-export$640/mo
- staging-runner$380/mo
- replay-worker$210/mo
Live agent · install in 5 minutes
Stop paying for unused capacity.
Connect your first cluster in 5 minutes. The first 100 vCPU is on us, forever. Above it, the rate tiers from $1.99 down to $0.60 — same features, every plan.
Self-hosted agentFirst 100 vCPU freeNo credit card