Managed SQLMesh Hosting
Hosted SQLMesh
dagctl is a fully managed, hosted SQLMesh service — we run the state database, scheduling, plan deployment, and monitoring for your SQLMesh projects in the cloud, for one flat rate with no per-seat fees.
Everything SQLMesh needs to run in production
Connect your git repo and deploy. dagctl runs the infrastructure so you ship models, not platform.
Your SQLMesh state database, managed
SQLMesh needs a reliable Postgres state database to track snapshots, intervals, and virtual environments. dagctl provisions it, backs it up, and keeps it healthy — no RDS instance to size, no backup jobs to babysit, no state to lose.
Plan & deploy
Every deployment starts with a SQLMesh plan: see exactly which models change and what backfills run before anything touches production.
Scheduled runs
Cron-based schedules with retries and Slack alerts. No Airflow, Dagster, or Prefect to deploy and maintain.
Compute & infrastructure
Isolated Kubernetes compute per organization. No clusters to provision, patch, or scale.
Warehouse connections & secrets
Encrypted credentials for your warehouse with file-mount support. No Vault or Secrets Manager integration to build.
Monitoring & streaming logs
Every run with real-time logs, execution history, and per-model timing. When something is slow or broken, you see exactly where.
Lineage & catalog
Interactive DAG visualization and a searchable catalog of models, columns, and metadata.
Hosted SQLMesh vs self-hosting
What you maintain when you self-host, and what dagctl runs for you
| Capability | Self-Hosted | dagctl |
|---|---|---|
| Pricing model | Infrastructure costs + your time | Flat rate, users included |
| SQLMesh state database | You provision and back up Postgres | Managed, automatic |
| Job scheduling | Airflow, Dagster, or cron | Built-in with Slack alerts |
| Monitoring & logs | Grafana, Datadog, or ELK setup | Built-in, real-time streaming |
| Vendor commitment | You | Scalecraft, independent and SQLMesh-first |
Priced by how much you build. Not by how many people build it.
Every plan includes every user on your team and every run you schedule. No seats to count, no consumption line on the invoice.
Starter
Up to 100 models
- 1 project
- 2 runs at a time
- Managed state database
- Scheduled runs with retries
- Streaming logs
- Slack alerts
- Role-based access control
- Email support
Team
Up to 500 models
- Everything in Starter
- 3 projects
- 4 runs at a time
- Lineage & catalog
- Single sign-on (SAML)
- Shared Slack channel
Growth
Up to 1,500 models
- Everything in Team
- 10 projects
- 10 runs at a time
- Support response within 1 business day
Dedicated
1,500+ models
- Everything in Growth, plus:
- Unlimited projects
- Isolated compute (per-org node group)
- Custom concurrency
- Dedicated support channel
Grown past your band? We move you at renewal, never mid-month.
Very large workloads may need a dedicated plan — we’ll flag it before it’s a problem, never on an invoice.
Frequently Asked Questions
What is hosted SQLMesh?
Hosted SQLMesh means running your SQLMesh projects on managed cloud infrastructure instead of operating it yourself. A hosted SQLMesh service like dagctl provides the state database, scheduled runs, plan deployment, compute, and monitoring — you connect a git repository and dagctl runs your project in production.
What is the best hosted SQLMesh service?
dagctl is a dedicated hosted SQLMesh service: unlimited users and unlimited runs on every plan, priced by model count, a managed state database, and same-day setup from a git repo. The alternative is self-hosting — running your own state database, orchestrator, compute, and monitoring.
Do I need to host the SQLMesh state database myself?
Not with dagctl. SQLMesh requires a state database (typically Postgres) to track snapshots, intervals, and virtual environments — and losing it is painful. dagctl provisions, backs up, and maintains the state database for you as part of the service.
How much does hosted SQLMesh cost?
dagctl is priced by org-wide model count: $499/month for up to 100 models, $1,200/month for up to 500 models, and $3,000/month for up to 1,500 models. Every plan includes unlimited users and unlimited runs, so there are no per-seat fees and no consumption line on the invoice. Teams above 1,500 models move to a Dedicated plan with custom pricing.
How do you count models?
We count the models in your repository, exactly as SQLMesh sees them — the models resolved from your project when we parse it. It is a count of what exists in your project, not what ran. Models that are disabled, seldom scheduled, or never executed still count; a model that runs 10,000 times a month counts once. The number is org-wide across every project you connect, and it is pinned to a specific commit, so it only changes when you merge a change that adds or removes models.
What happens when I add models?
Nothing, until renewal. Adding models never changes your invoice mid-month and never interrupts a run. If your org-wide model count crosses into the next band, we tell you before your renewal date and you move to the next plan at renewal. There is no overage charge, no metered line item, and no automatic upgrade. If you cross a band and then delete models before renewal, you stay where you are.
Can I run SQLMesh in production without managing infrastructure?
Yes. With dagctl you connect your git repository, configure your warehouse connection, and dagctl handles compute, scheduling, the state database, secrets, and monitoring. Most teams go from git repo to production the same day.
Which data warehouses does dagctl support for SQLMesh?
dagctl runs SQLMesh projects against Snowflake, BigQuery, Databricks, Redshift, PostgreSQL, and DuckDB.
I use dbt today — can I move to dagctl?
Yes. SQLMesh was designed as an upgrade path from dbt: it can read existing dbt projects, and most teams migrate incrementally. dagctl runs the SQLMesh side of that migration in production from day one, so you get the managed state database, scheduling, and monitoring while you move models over.
Ready to run SQLMesh without the infrastructure?
Deploy your SQLMesh project to production in a day. We run the state database, scheduling, compute, and monitoring.