Getting started /
Porting from Heroku

Porting from Heroku

Tapitalee gives you a Heroku-style workflow (build an image, deploy, attach add-ons, set config vars, run one-off commands) but the infrastructure lives in your own AWS account. Apps run as containers on ECS Fargate, databases are RDS instances, files go in S3, and you pay AWS directly for what you use.

If you already know Heroku, most concepts carry across one-to-one. This section explains the differences and the equivalent tapit commands so you can get an app moved with minimal surprises.

The big picture

Heroku Tapitalee Notes
Heroku account / Team Team Every user has a Personal team; create a team to collaborate. Teams are linked to one AWS account.
App App Belongs to a team, lives in one AWS region (in a VPC).
Dyno / process type (web, worker) Process The default process is the internet-facing one (Heroku web). Other processes are private background services.
Dyno size (Standard-1X, Performance-M) Process cpu / memory Any Fargate combination of vCPU and GB, set per process.
Dyno count (heroku ps:scale web=3) demand_count + spot_count Spot containers cost roughly a third of on-demand. Optional CPU autoscaling.
Slug / buildpack build Container image Built with Cloud Native Buildpacks (no Dockerfile needed) or your Dockerfile, pushed to a private ECR registry.
Release (v123) Deploy and optional Release record Rolling deploy with automatic rollback on failure.
Release phase (release: in Procfile) Pre-deploy steps Ordered commands run before each deploy completes; a failure aborts the deploy.
Config vars Variables Plain variables plus secrets stored in AWS Secrets Manager.
Add-ons (Postgres, Redis, etc.) Add-ons Real AWS services (RDS, ElastiCache, S3, EFS, …) in your account.
heroku run Tasks and console sessions One-off ECS tasks in a fresh ephemeral container.
Heroku Scheduler Scheduled commands Real cron expressions, not just 10-min/hourly/daily.
Custom domains + ACM / Automated Certificate Management Domains + SecureProxy Let’s Encrypt certificates via a Traefik proxy, or use Cloudflare.
heroku logs --tail tapit show logs -f CloudWatch Logs under the hood.
Review apps Preview apps Clone of the parent app, optional auto-delete after N days.
Pipelines / promote Multiple apps + create deploy docker_tag=... The same image tag can be deployed to any app in the team.
API key / HEROKU_API_KEY in CI Deploy tokens Scoped tokens for CI/CD.
Collaborators / Team members Memberships and roles read_only, member, admin, owner, plus granular access rights.

What is different

  • You bring the AWS account. Tapitalee provisions and manages resources in it. Billing for compute, databases and storage is on your AWS bill, not bundled.
  • Containers, not slugs. Every deploy is a Docker image. If you have no Dockerfile, tapit image deploy builds one for you with buildpacks, much like Heroku did.
  • One public process. Only the default process receives internet traffic. There is no router in front of your app unless you add a proxy add-on (SecureProxy, CertWrapper or Cloudflare), which is where TLS termination and zero-downtime switchover happen. See Webserving.
  • Dynamic IPs, no herokuapp.com router. Apps get an internal *.dns.tapitalee.net hostname to CNAME to. IPs change on deploy, so point DNS at the hostname or use the Route53 or Cloudflare add-ons to manage records for you.
  • No dyno sleeping, no 24-hour restarts. Containers keep running until you deploy, restart or disable them.
  • No 30-second router timeout, no 512 MB free tier. Resource limits are whatever you allocate.
  • Ephemeral filesystem, as on Heroku, but you can attach a persistent shared volume with the EFS add-on.
  • Regions. Heroku offered US and EU (plus Private Spaces). Tapitalee can deploy to any AWS region; each team gets one VPC per region.

Sections

Processes
Dynos, Procfile, scaling, release phase, one-off dynos and the Scheduler mapped to processes, commands and tasks.
 Add-ons
Heroku add-on equivalents, config var names, and common add-ons Tapitalee does not provide (static IPs, etc.).
 Databases
Heroku Postgres and Redis to RDS and ElastiCache: plans, backups, followers, credentials and data migration.
 Steps to porting an app
An ordered checklist from creating the team to cutting over DNS.

CLI at a glance

The tapit CLI follows a verb resource key=value pattern rather than Heroku’s topic:command pattern. Almost every command takes -t TEAM -a APP (or TAPIT_TEAM / TAPIT_APP env vars, or a tapit.toml created by tapit init).

heroku tapit
heroku login tapit ping (logs in if needed)
heroku apps tapit -t TEAM list apps
heroku create myapp tapit -t TEAM create app myapp region=us-east-1
heroku info tapit show app / tapit show summary
git push heroku main tapit image deploy
heroku releases tapit list deploys / tapit list releases
heroku rollback v122 tapit create deploy docker_tag=<previous tag>
heroku ps tapit list processes / tapit show service
heroku ps:scale web=2 worker=1 tapit set process name=default demand_count=2 etc.
heroku ps:restart tapit restart process name=default
heroku config tapit list variables / tapit export variables
heroku config:set KEY=val tapit set variable KEY val [secret=yes]
heroku config:unset KEY tapit delete variable KEY
heroku run rails console tapit run 'rails console'
heroku run:detached rake task tapit create task 'rake task'
heroku logs --tail tapit show logs -f
heroku addons tapit list addons
heroku addons:create heroku-postgresql tapit create rds engine=postgres name=maindb
heroku addons:create heroku-redis tapit create elasticache engine=redis name=cache
heroku addons:destroy NAME tapit delete addon name=NAME
heroku domains:add www.example.com tapit create domain name=www.example.com
heroku certs:auto:enable tapit create secureproxy acme_email=you@example.com
heroku pg:backups:capture tapit create snapshot addon=maindb
heroku pg:psql tapit run bash then psql $DATABASE_URL
heroku apps:destroy tapit delete app myapp

Run tapit help, tapit help <command> or browse the CLI and API reference for every option.