Tindra
Sign in Start free trial

One binary. One Postgres. Every error.

Error tracking and performance monitoring: errors, traces, profiles, logs, uptime and cron. Self‑host it in 30 seconds, or let us run a dedicated instance for you in the EU.

Start free trial
bash -c "$(curl -sSL https://install.tindra.sh)"

14 days free. No credit card. Or run it yourself.

Latest release v1.5.0 Released Release notes

Tindra dashboard with failing monitors, new and regressed issues, vitals, release health, recent alerts, slowest transactions, per-project metrics, busiest issues, request and latency charts and a seven-day traffic heatmap Tindra dashboard with failing monitors, new and regressed issues, vitals, release health, recent alerts, slowest transactions, per-project metrics, busiest issues, request and latency charts and a seven-day traffic heatmap
1 Go binary.
1 Postgres.
30s From install to running.
0 Queues, workers or sidecars to keep alive at 3am.

That is the whole stack.

No Kafka. No ClickHouse cluster. No separate worker processes. Pull the image, point it at Postgres, and it runs. Migrations happen on startup.

512 MB of RAM is the floor. 1 GB is comfortable.

On Debian or Ubuntu? There is an APT package.

Start at the error. End at the function.

  1. Error. The stack trace, the breadcrumbs, and the user it happened to.
  2. Logs. Their structured logs. Same time range, same filters.
  3. Trace. The request that broke, span by span.
  4. Profile. The sampled flame graph. Which function, not just which query.

Pick an affected user and Tindra keeps them in view. Project, environment, user and time filters follow you across issues, logs, traces and browser page loads. Share the link and a teammate lands in the same spot.

See the investigation workflow

Customer investigation with Jane Doe selected, her issues listed and the shared filters set Customer investigation with Jane Doe selected, her issues listed and the shared filters set
Jane Doe, followed across her issues. Real product, seeded demo data.

The whole list. It fits on one poster.

Errors
Stack traces, breadcrumbs and release tagging, grouped into issues you resolve or ignore. Source maps get checked against real frames from your build.
Performance
Transactions, span waterfalls, Apdex and Web Vitals. Flame graphs from your SDK’s profiler, switched on or off per project. Profiles never count as events.
Structured logs
Search messages, inspect attributes, open the matching trace. Hit Alert on this and a focused search becomes a volume alert. See logs and alerts
Uptime and cron
HTTP and HTTPS probes with down and recovery alerts. Cron monitors that notice a missed run or a failed one.
Alerts
Rules that deliver to email, Slack, Discord, Microsoft Teams or a webhook. Delivery history is kept, and failed webhooks retry on their own.
MCP server
Built into the binary. Ask your AI assistant about stack frames, breadcrumbs and older occurrences. Project-scoped tokens stay read-only until you say otherwise.
Your own instance
Managed customers get a dedicated container and a dedicated Postgres. Not a row in a shared table, not a schema-per-tenant trick.
Sentry SDKs
Already on a Sentry SDK? Keep it and change the DSN. Guided setup waits until a test event is actually stored. Get your app connected

Every feature, with screenshots

The waterfall says which query. The flame graph says which function.

Turn on your SDK’s profiler and profiles attach to sampled transactions. Open the transaction and the flame graph sits right under the spans.

How profiling works

One profiled transaction, GET /api/orders/export, with its summary, span waterfall and sampled flame graph One profiled transaction, GET /api/orders/export, with its summary, span waterfall and sampled flame graph
GET /api/orders/export. Summary, waterfall and flame graph of one profiled transaction. Seeded demo data.

Run it yourself. Or don’t.

Same product, same UI, same SDK protocol. The only difference is whose pager goes off at 3am.

$0

Self-hosted

Free and source available under the Elastic License 2.0. Docker image or a native Debian and Ubuntu package.

bash -c "$(curl -sSL https://install.tindra.sh)"

The installer asks a few questions, writes a docker-compose.yml and starts the stack. Prefer your own compose file? Then it is docker compose up -d.

  • No telemetry, no license server
  • No event limits
  • Same UI and SDK protocol as managed
  • Support through GitHub Issues and Discussions
Read the install guide View on GitHub

$9/mo

Managed in the EU

Sign up, pick a hostname, and your own instance is live in a few minutes. Backups, upgrades, retention and the pager are on us.

  • Dedicated container and Postgres per customer
  • Hosted in the EU, never replicated outside
  • Nightly encrypted backups, 14-day point-in-time recovery
  • OAuth and OIDC single sign-on on Studio and Scale
  • Signed DPA, GDPR Art. 28
  • 99.9% uptime SLA on Studio and Scale
Start free trial See pricing

Your data stays in the EU. Not for backups. Not even by mistake.

Managed instances run entirely in the EU. Backups replicate to a second availability zone in the same region. Object storage stays EU‑only. No US sub-processors.

Residency
EU only
Backups
14-day point-in-time recovery
DPA
GDPR Art. 28, signed
Deletion
Within 24 hours

Operated by Blendbyte GmbH, Berlin. We use Tindra in production ourselves.

Pick a number. It won’t change on you.

Every plan has a monthly event limit. Events are errors plus transactions. Logs, cron check-ins and profiles don’t count. Hit the limit and ingestion pauses until the next cycle. No overage, no auto-billing. Need more? Upgrade the plan.

Hobby

One app, up to three people.

$9/mo

Billed monthly

50,000 events / mo
30 days retention

  • 1 project
  • 3 user accounts
  • 500K logs, 1M transaction rows, 256 MB profiles
  • Email support
Start free trial

Indie

Every app you run, the whole team.

$39/mo

Billed monthly

500,000 events / mo
60 days retention

  • Unlimited projects
  • Unlimited users
  • 10M logs, 25M transaction rows, 1 GB profiles
  • Email support
Start free trial

Studio Recommended

Busy production apps that need SSO and an SLA.

$199/mo

Billed monthly

5,000,000 events / mo
90 days retention

  • Unlimited projects and users
  • 50M logs, 100M transaction rows, 4 GB profiles
  • SSO and OIDC
  • 99.9% uptime SLA
  • Priority email support
Start free trial

Scale

High volume, an audit log and a named contact.

$549/mo

Billed monthly

50,000,000 events / mo
90 days retention

  • Unlimited projects and users
  • 250M logs, 500M transaction rows, 8 GB profiles
  • SSO, OIDC, audit log
  • 99.9% uptime SLA
  • Dedicated account manager
Start free trial

$0

Self-hosted is free. No event limits, same product. Elastic License 2.0.

Self-hosting guide
  • Log, transaction and profile storage limits are instance-wide totals. The oldest entries go first when full.
  • Every plan: 14-day free trial, no credit card.
  • Prices are net, excluding VAT.

One binary and Postgres. Any box, 512 MB.

How the essentials line up. Checked against each vendor’s public pages in July 2026.

Capability Tindra Sentry GlitchTip AppSignal
Performance monitoringYesYesYesYes
Profiling and flame graphsYesYesNoNo
Log viewerYesYesNoNo
Uptime monitoringYesYesYesYes
Cron monitoringYesYesYesYes
MCP supportBuilt inYes*NoNo
Entry plan$9 / moFree (1 user) / $29 / moFree (1K events) / $15 / mo$23 / mo
Self-hostingBinary + Postgres, 512 MB floorFree, ~65 containers by default, 16 GB RAM minFree, Django, Redis, Celery and PostgresNot available
EU data residencyAll plansOptional add-onNot availableYes, by default
At the limitEvents pause, no auto-billingAuto-billed per eventThrottled, blocked at 2x quotaNo hard limits

* Sentry’s MCP server is a separate hosted service, or a local process when self-hosted. Tindra’s is built into the binary. Verified July 2026.

Pick a box. Or pick a plan.

Start free trial Read the docs
bash -c "$(curl -sSL https://install.tindra.sh)"

Managed: live in a few minutes, 14 days free. Self‑hosted: one command, running before your coffee is done.