Updog
Prod did a thing. Find out wuz Updog.

Stop guessing and start figuring out WTF is updog.

Errors, logs, metrics, and performance data in one investigation trail—so your team can stop opening tabs and start closing incidents.

Let Updog investigate Log in
One connected incident trail Portal and CLI workflows Project-scoped access

THE CONTEXT THAT MATTERS, IN ONE PLACE

A dashboard with receipts.

Updog keeps the error, the noisy logs, the suspicious metric, and the slow request in the same frame. No detective corkboard required.

Checkout Platform

Project telemetry and investigation workspace

Monitoring active

UNRESOLVED INCIDENT

Unhandled payment exception

18 occurrences · last seen 42 seconds ago
3
open groups
18
occurrences
Investigate error

Recent errors

Newest signals from this project

View all
LAST SEENSTATUSERRORCOUNT
UnresolvedUnhandledPaymentException18
ResolvedDatabaseConnectionError7
IgnoredTimeoutError31

Signal trail

Everything changed at 14:02 UTC

30m
Errors+18
Matching logs342
Error rate+892%

Jump to investigate

Move from the signal into the supporting telemetry.
Logs Metrics Performance Errors

PRODUCTION CONTEXT FOR CODING AGENTS

Let your agent investigate without handing it the keys.

Updog gives coding agents a project-scoped, read-only path into current errors and logs. They can search, pull full error context, and bring the evidence back to the code—without deploy credentials or write access.

  • Read-only by design Search and inspect production telemetry. Never mutate it.
  • Made for agent workflows Human-readable output or predictable JSON when the agent wants structure.
  • Enough context to be useful Search recent signals, then retrieve the complete error trail.
Codex checkout-platform
READ-ONLY
YOU

Why is production down?!

CODEX INVESTIGATING

I’ll check the latest project-scoped signals in Updog before I start guessing at the code.

UPDOG CLI 2 READ-ONLY QUERIES
$ updog --project checkout --json errors search --query "payment" --since 1h

18 errors · UnhandledPaymentException · first seen 14:02 UTC

$ updog --project checkout --json logs search --query "DBConnection.ConnectionError" --since 1h

342 matching logs · 12 instances · deploy 8f32a1

Found it. checkout-api exhausted its database pool immediately after deploy 8f32a1. The spike began at 14:02 UTC across 12 instances.

Project-scoped access Evidence attached

LESS TAB MIGRATION. MORE INCIDENT RESOLUTION.

One investigation. All the context.

Follow the first signal into the logs, metrics, traces, and deployment context behind it—without losing the thread.

01 · AGENT-FRIENDLY CLI

Give your agents production context—not production access.

Updog’s read-only CLI lets coding agents search project-scoped errors and logs, then pull the full error context when the evidence gets interesting.

$ updog --project checkout errors show err_8421

02 · KEEP THE THREAD

Errors, logs, metrics, and performance stop pretending they have never met.

  • Project-scoped telemetry
  • Searchable incident context

03 · SHARE THE ANSWER

Give the team one investigation instead of a group chat full of screenshots.

Seven teammates. One version of what happened.

THE INCIDENT FLOW

Alert. Evidence. Answer. Nap.

  1. 01CaptureThe production signal arrives.
  2. 02ConnectRelated telemetry stays together.
  3. 03InvestigateFollow the useful evidence.
  4. 04ResolveKnow what happened and why.

READY WHEN PRODUCTION ISN'T

Production down? Find out wuz Updog.

$ updog investigate --incident INC-8421