inkvo
Log inStart free
NEWNarrative incidents, powered by context from your stack

The first SRE on your team
— even if you don't have one.

When something breaks at 3am, Inkvo doesn't dump graphs on you. It reads the signals, correlates the deploys, and tells you what happened — in plain English.

Start free
No credit card · 14-day Pro trial · Connect OTel in 4 minutes
app.inkvo.dev / incidents / inc-2847
SEV-1prodinc-2847
Checkout timeouts on /api/orders
opened 03:12:41 UTC · 8m ago
Affected services
● payments-svc ● orders-api + 2 downstream
Customer impact
~ 2 failed checkouts / min
Correlated deploy
payments-svc @ 3d9a1c · 03:08 UTC
Paged
@maya (you) · 03:12
NARRATIVE · URGENT

At 03:12 UTC, checkout p99 latency went from 180ms → 0.2s and started returning 504s. The root cause is payments-svc: its outbound calls to Stripe are timing out at the connection layer.

Four minutes earlier, payments-svc shipped 3d9a1c — a dependency bump that included stripe-node 14 → 15. The new client defaults to IPv6 and your egress NAT doesn't route IPv6 to the internet.

Suggested: roll back payments-svc to 5f20e8 (takes ~90s). A fix-forward would pin the client to IPv4 — I drafted it as PR #2104.

$ inkvo rollback payments-svc --to 5f20e8
↓ waiting for ack…
On-call for teams who ship
mercuriaparallelarcfieldsnowcapobol labsdelta·vquillhalitemeridianphosphormercuriaparallelarcfieldsnowcapobol labsdelta·vquillhalitemeridianphosphor
The 3am problem

You get paged.
You open four dashboards.
You have no idea where to start.

03:12
Page fires
PagerDuty wakes you up. "Checkout error rate ≥ 5%."
03:13
Open Datadog
You stare at 17 graphs. Something spiked. You can't tell which thing is the cause.
03:16
Open Grafana
Different dashboard, different time range. p99 is red. Why?
03:18
Open GitHub
Deploys… there were three. You start reading diffs on your phone.
03:24
Open logs
Stream of 500s. You grep for "stripe." You find nothing useful.
03:31
Wake up coworker
Because you're 19 minutes in and still don't know what broke.
The industry's answer to this is "train an SRE." Most teams of 5–50 don't have one.
How it works

Three steps. Afternoon of setup, then quiet nights.

01Connect
Point your OTel collector at us.
One endpoint, one API key per environment. No sidecars, no new agents. Inkvo ingests traces, metrics, and logs in whatever shape you already send them.
OTEL_EXPORTER_OTLP_ENDPOINT=ingest.inkvo.dev
OTEL_EXPORTER_OTLP_HEADERS="authorization=ik_prod_b4f2…"
02Detect
We learn what normal looks like.
After 24 hours, Inkvo knows your baseline per service, per environment. Thresholds adapt with your traffic — strict in prod, relaxed in staging, debug-only in dev.
03Explain
When something breaks, you get a story.
Not a graph. A narrative: what broke, what's affected, what changed right before it broke, and a suggested next step. You read it in 20 seconds and know what to do.
Tone adapts to environment

A prod outage reads like war.
A staging hiccup reads like a heads-up.

NARRATIVE · URGENT
Checkout is down in prod. ~34 failed orders/min. payments-svc can't reach Stripe — the 03:08 deploy bumped stripe-node to v15, which defaults to IPv6, and your NAT doesn't route IPv6 out. Suggested: roll back to 5f20e8 now.
One feed, three jobs

On-call hits everyone. Inkvo speaks everyone's language.

Your reliability work, leveraged.
Stop being the only person who can read the dashboards. Inkvo turns your team's tribal knowledge into narratives anyone on-call can follow — and you get a clean audit trail when the post-mortem happens.
  • Custom signal weights per service
  • Runbook attach + suggest-on-incident
  • SLO burn-rate alerts with narrative context
  • Post-mortem export to Markdown
SLO BURN · payments-svclast 30d
Error budget · 14% remaining
RUNBOOK · attach
↳ payments-svc / stripe-timeout.md ✓ linked
↳ infra / egress-nat-failover.md ✓ linked
Plays well with what you have

One endpoint in. Every signal you already collect.

Inkvo doesn't replace your stack. It reads from it. OTel is the fast path; everything else is a tile away.

Telemetry
O
OpenTelemetry
OTLP ingest
P
Prometheus
metrics pull
G
Grafana
dashboards
H
Honeycomb
traces
D
Datadog
forward
Cloud & Infra
A
AWS
CloudWatch · ECS · Lambda
G
GCP
Cloud Logging · GKE
A
Azure
Monitor · AKS
K
Kubernetes
events · logs
C
Cloudflare
workers · logs
Source & Deploys
G
GitHub
PRs · deploys
G
GitLab
CI · MRs
V
Vercel
deploys
N
Netlify
deploys
C
CircleCI
pipelines
Paging & Comms
P
PagerDuty
rotations · alerts
O
Opsgenie
rotations
S
Slack
channels · DMs
MT
MS Teams
channels
L
Linear
auto-create issues
Don't see your stack? Inkvo speaks raw OTLP, JSON over HTTP, and webhooks.Browse all 40+ connectors →
vs. the legacy stack

Built narrative-first. Priced for teams, not enterprises.

Feature
Datadog
PagerDuty
inkvo
Setup time
Days. Agents, collectors, custom configs.
Days. Per-service integrations.
Afternoon. One OTel endpoint.
What you see at 3am
17 dashboards to stare at.
An alert. Go find the cause.
A narrative. Root cause already guessed.
Environments
Tags you curate forever.
Services × policies × escalations.
Built-in. Prod/staging/dev with sane defaults.
Pricing
Per host, per GB, per custom metric.
Per seat, per integration.
Per service monitored. That's it.
Built for
Enterprises with SRE teams.
Enterprises with SRE teams.
Teams of 5–50 without one.
Pricing

Priced per service. Not per byte, seat, or panic.

Starter
For solo devs and tiny teams. Get paged less.
Free
up to 5 services
  • 5 services
  • 2 environments
  • Narrative incidents
  • Slack + email
  • 14-day history
Start free
MOST TEAMS
Team
Most teams of 5–50 land here.
$24/mo
per service / month
  • Unlimited environments
  • Unlimited seats
  • PagerDuty + Opsgenie
  • On-call rotations
  • 90-day history
  • Draft rollback PRs
Start 14-day trial
Business
For teams with compliance or volume needs.
Custom
talk to us
  • SOC 2, HIPAA BAA
  • SSO + SCIM
  • Custom retention
  • Audit log export
  • Dedicated support
  • Self-hosted option

Sleep better on-call.

Connect your first service in an afternoon. The next time you're paged, you'll open one tab, not four.

Start free
all systems nominal·ingest p99 148ms·regions us-east · us-west · eu-weststatus.inkvo.dev →