Redash Review: Self-Hosted SQL Dashboards, Security & Status
What it is: Redash is an open-source, SQL-first analytics platform: write queries against 50+ data sources, visualize the results, and assemble dashboards your team can share. It's self-hosted only (the cloud service was discontinued), lightweight, and built for teams that live in SQL.
Quick answer: Redash is still the simplest free way to turn SQL queries into shared dashboards, and self-hosting it with Docker takes under an hour. The honest caveats: development slowed sharply after Databricks acquired it in 2020 (community releases have since resumed), and securing a deployment — SSO, group permissions, network isolation — is entirely on you. For non-technical users, pick Metabase instead.
How Do You Secure a Redash Deployment?
Securing Redash means layering four controls: keep the instance off the public internet (VPN or reverse proxy with HTTPS), enforce strong authentication (Google OAuth or SAML SSO instead of passwords), scope database access with read-only credentials per data source, and use Redash groups to control which users can see which data sources. Since you self-host, none of this is on by default — it's your checklist, not the vendor's.
Concrete hardening steps that catch most real-world problems:
Network: Terminate TLS at a reverse proxy (nginx/Caddy), set REDASH_ENFORCE_HTTPS, and restrict access via VPN, IP allowlist, or an auth proxy — publicly exposed Redash instances are a recurring source of data leaks.
Authentication: Enable Google OAuth or SAML, disable password login where possible, and require admin approval for new accounts.
Database credentials: Connect every data source with a read-only user scoped to only the schemas analysts need. Redash executes raw SQL, so the database user's grants are your real permission boundary.
Groups and sharing: Map data sources to Redash groups (e.g. finance data → finance group), audit who holds admin, and be deliberate with public sharing links — anyone with the URL can see a publicly shared dashboard.
Secrets: Keep REDASH_COOKIE_SECRET and REDASH_SECRET_KEY out of source control, and update images regularly for security patches.
Is Redash Still Maintained After the Databricks Acquisition?
Partially — be clear-eyed about this. Databricks acquired Redash in June 2020, the hosted cloud product shut down in 2021, and official development largely stalled after v10 (October 2021). The project was handed to community maintainers, who have since shipped new releases (the 25.x series), but the pace remains slower than Metabase or Superset. Redash works fine today; just don't expect rapid feature development, and have a migration story in mind.
What It Does Best
SQL-focused. No drag-and-drop abstraction. You write SQL, it visualizes results. Perfect for analysts.
Query library. Save and share queries. Fork others' queries. Collaboration built-in.
Scheduled queries. Run queries on schedule, email results. Simple alerts to Slack or email when values cross thresholds.
Key Features
SQL editor: Write queries with autocomplete, visualize results instantly
50+ data sources: PostgreSQL, MySQL, BigQuery, Redshift, MongoDB, Elasticsearch, and more
Query library: Browse, fork, and share queries across the team
Query parameters: Dropdowns and date pickers turn queries into lightweight apps
Scheduled refreshes: Auto-update dashboards on schedule
Alerts: Email/Slack notifications when values change
API: Programmatic access to queries and dashboards
Pricing
Open source: Free (BSD-2 license), self-hosted
Real cost: A small VM (2 vCPU / 4–8 GB RAM runs it comfortably) plus your ops time for upgrades, backups, and the security hardening above
Cloud: Discontinued in 2021 — self-hosting is the only official option
How Do You Self-Host Redash?
The standard route is Docker Compose: the community setup script provisions the Redash server, worker, Redis, and PostgreSQL containers on a single VM in under an hour. From there, put a TLS-terminating reverse proxy in front, configure SSO, and schedule Postgres backups. One-click images also exist on the AWS and DigitalOcean marketplaces if you want to skip the manual setup.
When to Use It
✅ Team writes SQL comfortably
✅ Simple visualization needs
✅ Want query sharing/collaboration
✅ Zero software budget, some ops capacity
✅ Need lightweight, fast setup
When NOT to Use It
❌ Non-technical users (need a no-code tool — see Metabase)
❌ Complex visualizations (limited chart types)
❌ Want a managed service (cloud offering ended)
❌ Need a fast-moving, actively developed platform
❌ Require enterprise features like row-level security out of the box
Common Use Cases
Data team dashboards: Quick SQL-based reporting for analysts
Operational monitoring: Scheduled queries for system health
Query collaboration: Share and reuse SQL across the team
Simple BI: Lightweight alternative to complex BI platforms
Alerts and monitoring: Automated notifications for data changes
Redash vs Alternatives
vs Metabase: Metabase is friendlier for non-technical users (visual question builder, better admin UX) and far more actively developed, with a managed cloud option; Redash stays cleaner for SQL-first teams. For most new deployments in 2026, Metabase is the safer default.
vs Superset: Superset has many more chart types, row-level security, and an active Apache community — but it's heavier to run and steeper to learn. Redash wins on simplicity; Superset wins on power.
vs Mode: Mode adds Python notebooks, versioning, and polish as a commercial product; Redash is free and self-hosted. Budget decides.
Unique Strengths
Simplicity: Easiest SQL-to-dashboard tool to set up
Query library: Best query sharing and forking experience
Lightweight: Minimal overhead, runs on a small VM
Open source: Free, self-hosted, no per-seat pricing ever
Bottom line: Still the simplest SQL-to-dashboard tool, and free forever if you self-host — just budget real effort for security (SSO, read-only credentials, network isolation) because that's entirely on you. Development has slowed since the Databricks acquisition, so evaluate Metabase or Superset for greenfield projects; choose Redash when SQL-first simplicity is exactly what you want.